{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "import seaborn as sns\n", "\n", "sns.set_style(\"whitegrid\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Reading in our potential test sets from [here](https://huggingface.co/datasets/imageomics/lila-bc-camera/tree/d18307b285217d18b31d1a7b2c9091bb0873ade0/data/potential-test-sets):\n", " - [Ohio Small Animals](https://lila.science/datasets/ohio-small-animals/)\n", " - [Desert Lion Conservation Camera Traps](https://lila.science/datasets/desert-lion-conservation-camera-traps/)\n", " - [Orinoquia Camera Traps](https://lila.science/datasets/orinoquia-camera-traps/)\n", " - [Island Conservation Camera Traps](https://lila.science/datasets/island-conservation-camera-traps/)\n", " - [ENA24](https://lila.science/datasets/ena24detection)\n", "\n", "We'll clean them down to just the taxa and identifier columns, then further reduce to make balanced test sets for each.\n", "\n", "# Desert Lion Conservation Camera Traps Datasets\n", "upper/lower bounded and balanced" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
dataset_nameurl_gcpurl_awsurl_azureimage_idsequence_idlocation_idframe_numoriginal_labelscientific_name...superfamilyfamilysubfamilytribegenusspeciessubspeciesvarietymulti_speciesnum_species
0Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...NaNfelidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0
1Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...NaNfelidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0
2Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...NaNfelidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0
3Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...NaNfelidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0
4Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...NaNfelidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0
\n", "

5 rows × 34 columns

\n", "
" ], "text/plain": [ " dataset_name \\\n", "0 Desert Lion Conservation Camera Traps \n", "1 Desert Lion Conservation Camera Traps \n", "2 Desert Lion Conservation Camera Traps \n", "3 Desert Lion Conservation Camera Traps \n", "4 Desert Lion Conservation Camera Traps \n", "\n", " url_gcp \\\n", "0 https://storage.googleapis.com/public-datasets... \n", "1 https://storage.googleapis.com/public-datasets... \n", "2 https://storage.googleapis.com/public-datasets... \n", "3 https://storage.googleapis.com/public-datasets... \n", "4 https://storage.googleapis.com/public-datasets... \n", "\n", " url_aws \\\n", "0 http://us-west-2.opendata.source.coop.s3.amazo... \n", "1 http://us-west-2.opendata.source.coop.s3.amazo... \n", "2 http://us-west-2.opendata.source.coop.s3.amazo... \n", "3 http://us-west-2.opendata.source.coop.s3.amazo... \n", "4 http://us-west-2.opendata.source.coop.s3.amazo... \n", "\n", " url_azure \\\n", "0 https://lilawildlife.blob.core.windows.net/lil... \n", "1 https://lilawildlife.blob.core.windows.net/lil... \n", "2 https://lilawildlife.blob.core.windows.net/lil... \n", "3 https://lilawildlife.blob.core.windows.net/lil... \n", "4 https://lilawildlife.blob.core.windows.net/lil... \n", "\n", " image_id \\\n", "0 Desert Lion Conservation Camera Traps : acinon... \n", "1 Desert Lion Conservation Camera Traps : acinon... \n", "2 Desert Lion Conservation Camera Traps : acinon... \n", "3 Desert Lion Conservation Camera Traps : acinon... \n", "4 Desert Lion Conservation Camera Traps : acinon... \n", "\n", " sequence_id \\\n", "0 Desert Lion Conservation Camera Traps : unknown \n", "1 Desert Lion Conservation Camera Traps : unknown \n", "2 Desert Lion Conservation Camera Traps : unknown \n", "3 Desert Lion Conservation Camera Traps : unknown \n", "4 Desert Lion Conservation Camera Traps : unknown \n", "\n", " location_id frame_num \\\n", "0 Desert Lion Conservation Camera Traps : unknown -1 \n", "1 Desert Lion Conservation Camera Traps : unknown -1 \n", "2 Desert Lion Conservation Camera Traps : unknown -1 \n", "3 Desert Lion Conservation Camera Traps : unknown -1 \n", "4 Desert Lion Conservation Camera Traps : unknown -1 \n", "\n", " original_label scientific_name ... superfamily family subfamily \\\n", "0 acinonyx jubatus acinonyx jubatus ... NaN felidae felinae \n", "1 acinonyx jubatus acinonyx jubatus ... NaN felidae felinae \n", "2 acinonyx jubatus acinonyx jubatus ... NaN felidae felinae \n", "3 acinonyx jubatus acinonyx jubatus ... NaN felidae felinae \n", "4 acinonyx jubatus acinonyx jubatus ... NaN felidae felinae \n", "\n", " tribe genus species subspecies variety multi_species \\\n", "0 NaN acinonyx acinonyx jubatus NaN NaN False \n", "1 NaN acinonyx acinonyx jubatus NaN NaN False \n", "2 NaN acinonyx acinonyx jubatus NaN NaN False \n", "3 NaN acinonyx acinonyx jubatus NaN NaN False \n", "4 NaN acinonyx acinonyx jubatus NaN NaN False \n", "\n", " num_species \n", "0 1.0 \n", "1 1.0 \n", "2 1.0 \n", "3 1.0 \n", "4 1.0 \n", "\n", "[5 rows x 34 columns]" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df = pd.read_csv(\"../data/potential-test-sets/Desert_Lion_Conservation_Camera_Traps_image_urls_and_labels.csv\", low_memory = False)\n", "df.head()" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Index(['dataset_name', 'url_gcp', 'url_aws', 'url_azure', 'image_id',\n", " 'sequence_id', 'location_id', 'frame_num', 'original_label',\n", " 'scientific_name', 'common_name', 'datetime', 'annotation_level',\n", " 'kingdom', 'phylum', 'subphylum', 'superclass', 'class', 'subclass',\n", " 'infraclass', 'superorder', 'order', 'suborder', 'infraorder',\n", " 'superfamily', 'family', 'subfamily', 'tribe', 'genus', 'species',\n", " 'subspecies', 'variety', 'multi_species', 'num_species'],\n", " dtype='object')" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.columns" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Observe that we also now get multiple URL options; `url_aws` will likely be best/fastest for use with [`distributed-downloader`](https://github.com/Imageomics/distributed-downloader) to get the images." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "RangeIndex: 63468 entries, 0 to 63467\n", "Data columns (total 34 columns):\n", " # Column Non-Null Count Dtype \n", "--- ------ -------------- ----- \n", " 0 dataset_name 63468 non-null object \n", " 1 url_gcp 63468 non-null object \n", " 2 url_aws 63468 non-null object \n", " 3 url_azure 63468 non-null object \n", " 4 image_id 63468 non-null object \n", " 5 sequence_id 63468 non-null object \n", " 6 location_id 63468 non-null object \n", " 7 frame_num 63468 non-null int64 \n", " 8 original_label 63468 non-null object \n", " 9 scientific_name 63468 non-null object \n", " 10 common_name 63468 non-null object \n", " 11 datetime 0 non-null float64\n", " 12 annotation_level 63468 non-null object \n", " 13 kingdom 63468 non-null object \n", " 14 phylum 63468 non-null object \n", " 15 subphylum 63468 non-null object \n", " 16 superclass 0 non-null float64\n", " 17 class 63468 non-null object \n", " 18 subclass 53378 non-null object \n", " 19 infraclass 53378 non-null object \n", " 20 superorder 53378 non-null object \n", " 21 order 63083 non-null object \n", " 22 suborder 24045 non-null object \n", " 23 infraorder 800 non-null object \n", " 24 superfamily 0 non-null float64\n", " 25 family 63037 non-null object \n", " 26 subfamily 32225 non-null object \n", " 27 tribe 20566 non-null object \n", " 28 genus 61117 non-null object \n", " 29 species 61040 non-null object \n", " 30 subspecies 11687 non-null object \n", " 31 variety 0 non-null float64\n", " 32 multi_species 63468 non-null bool \n", " 33 num_species 63468 non-null float64\n", "dtypes: bool(1), float64(5), int64(1), object(27)\n", "memory usage: 16.0+ MB\n" ] } ], "source": [ "df.info(show_counts = True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This is mostly filled in. All are labeled down to class level. Let's get some counts on the taxa, then start looking at balancing it." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "lin_taxa = ['kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']\n", "taxa_cols = ['original_label', 'scientific_name', 'common_name', 'kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "original_label 46\n", "scientific_name 45\n", "common_name 46\n", "kingdom 1\n", "phylum 1\n", "class 2\n", "order 17\n", "family 24\n", "genus 35\n", "species 38\n", "dtype: int64" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df[taxa_cols].nunique()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We only have 45--but there's one more `original_label` and `common_name`, but then only 38 different species. This is maybe since they're all labeled to `class`, but about 2K not beyond to `species`.\n", "\n", "Also, should check for duplicated images." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "number of unique images: 63468\n" ] }, { "data": { "text/plain": [ "multi_species\n", "False 63468\n", "Name: count, dtype: int64" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "print(f\"number of unique images: {df[\"image_id\"].nunique()}\")\n", "\n", "df[\"multi_species\"].value_counts()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Okay, none of these have multiple species per image, so now let's look at how many images we have per species." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
dataset_nameurl_gcpurl_awsurl_azureimage_idsequence_idlocation_idframe_numoriginal_labelscientific_name...superfamilyfamilysubfamilytribegenusspeciessubspeciesvarietymulti_speciesnum_species
55164Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : pteroc...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1pteroclidaepteroclidae...NaNpteroclidaeNaNNaNNaNNaNNaNNaNFalse1.0
55169Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : pteroc...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1pteroclidaepteroclidae...NaNpteroclidaeNaNNaNNaNNaNNaNNaNFalse1.0
5772Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : cn-rap...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1cn-raptorsaves...NaNNaNNaNNaNNaNNaNNaNNaNFalse1.0
55207Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : pteroc...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1pteroclidaepteroclidae...NaNpteroclidaeNaNNaNNaNNaNNaNNaNFalse1.0
\n", "

4 rows × 34 columns

\n", "
" ], "text/plain": [ " dataset_name \\\n", "55164 Desert Lion Conservation Camera Traps \n", "55169 Desert Lion Conservation Camera Traps \n", "5772 Desert Lion Conservation Camera Traps \n", "55207 Desert Lion Conservation Camera Traps \n", "\n", " url_gcp \\\n", "55164 https://storage.googleapis.com/public-datasets... \n", "55169 https://storage.googleapis.com/public-datasets... \n", "5772 https://storage.googleapis.com/public-datasets... \n", "55207 https://storage.googleapis.com/public-datasets... \n", "\n", " url_aws \\\n", "55164 http://us-west-2.opendata.source.coop.s3.amazo... \n", "55169 http://us-west-2.opendata.source.coop.s3.amazo... \n", "5772 http://us-west-2.opendata.source.coop.s3.amazo... \n", "55207 http://us-west-2.opendata.source.coop.s3.amazo... \n", "\n", " url_azure \\\n", "55164 https://lilawildlife.blob.core.windows.net/lil... \n", "55169 https://lilawildlife.blob.core.windows.net/lil... \n", "5772 https://lilawildlife.blob.core.windows.net/lil... \n", "55207 https://lilawildlife.blob.core.windows.net/lil... \n", "\n", " image_id \\\n", "55164 Desert Lion Conservation Camera Traps : pteroc... \n", "55169 Desert Lion Conservation Camera Traps : pteroc... \n", "5772 Desert Lion Conservation Camera Traps : cn-rap... \n", "55207 Desert Lion Conservation Camera Traps : pteroc... \n", "\n", " sequence_id \\\n", "55164 Desert Lion Conservation Camera Traps : unknown \n", "55169 Desert Lion Conservation Camera Traps : unknown \n", "5772 Desert Lion Conservation Camera Traps : unknown \n", "55207 Desert Lion Conservation Camera Traps : unknown \n", "\n", " location_id frame_num \\\n", "55164 Desert Lion Conservation Camera Traps : unknown -1 \n", "55169 Desert Lion Conservation Camera Traps : unknown -1 \n", "5772 Desert Lion Conservation Camera Traps : unknown -1 \n", "55207 Desert Lion Conservation Camera Traps : unknown -1 \n", "\n", " original_label scientific_name ... superfamily family subfamily \\\n", "55164 pteroclidae pteroclidae ... NaN pteroclidae NaN \n", "55169 pteroclidae pteroclidae ... NaN pteroclidae NaN \n", "5772 cn-raptors aves ... NaN NaN NaN \n", "55207 pteroclidae pteroclidae ... NaN pteroclidae NaN \n", "\n", " tribe genus species subspecies variety multi_species num_species \n", "55164 NaN NaN NaN NaN NaN False 1.0 \n", "55169 NaN NaN NaN NaN NaN False 1.0 \n", "5772 NaN NaN NaN NaN NaN False 1.0 \n", "55207 NaN NaN NaN NaN NaN False 1.0 \n", "\n", "[4 rows x 34 columns]" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna()].sample(4)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "class\n", "aves 2415\n", "mammalia 13\n", "Name: count, dtype: int64" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna(), \"class\"].value_counts()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Ahh, a lot of birds and some mammals. These are the only two classes in this dataset. \n", "\n", "We probably still want to restrict down to species. Let's see where the remaining different scientifc names are from, perhaps order." ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "order\n", "pterocliformes 1783\n", "galliformes 143\n", "columbiformes 71\n", "strigiformes 33\n", "carnivora 13\n", "Name: count, dtype: int64" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna(), \"order\"].value_counts()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Yes, those labeled only to `order`, though we also may have some families and genera without species too." ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "family\n", "pteroclidae 1783\n", "phasianidae 77\n", "columbidae 71\n", "numididae 66\n", "Name: count, dtype: int64" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna(), \"family\"].value_counts()" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "genus\n", "francolinus 77\n", "Name: count, dtype: int64" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna(), \"genus\"].value_counts()" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "scientific_name\n", "pteroclidae 1783\n", "aves 385\n", "francolinus 77\n", "columbidae 71\n", "numididae 66\n", "strigiformes 33\n", "carnivora 13\n", "Name: count, dtype: int64" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.loc[df[\"species\"].isna(), \"scientific_name\"].value_counts()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Probably still want to limit to just those with species-level designations." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "scientific_name\n", "oryx gazella 16499\n", "equus zebra hartmannae 11687\n", "parahyaena brunnea 7814\n", "struthio camelus 6076\n", "antidorcas marsupialis 3457\n", "giraffa camelopardalis 2935\n", "diceros bicornis 2736\n", "pteroclidae 1783\n", "panthera leo 1730\n", "lupulella mesomelas 1462\n", "loxodonta africana 1434\n", "crocuta crocuta 1121\n", "torgos tracheliotos 709\n", "hystrix africaeaustralis 544\n", "corvus albus 511\n", "Name: count, dtype: int64" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.scientific_name.value_counts()[:15]" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "scientific_name\n", "eupodotis rueppelii 35\n", "equus asinus 34\n", "strigiformes 33\n", "corvus capensis 26\n", "oreotragus oreotragus 25\n", "otocyon megalotis 22\n", "carnivora 13\n", "proteles cristatus 12\n", "bos taurus 11\n", "herpestes sanguineus 10\n", "alopochen aegyptiaca 7\n", "felis catus 7\n", "felis lybica 5\n", "genetta genetta 3\n", "gyps africanus 3\n", "Name: count, dtype: int64" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.scientific_name.value_counts()[30:]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Could potentially set the lower bound on number of images a bit higher, but let's look at those least-represented species:\n", " - _gyps africanus_ critically endangered (population decreasing) [IUCN](https://www.iucnredlist.org/species/22695189/204461164), but also most common African vulture [Wikipedia](https://en.wikipedia.org/wiki/White-backed_vulture) ([source article](https://africageographic.com/stories/vultures/)).\n", " - _genetta genetta_ listed as stable, least concern by [IUCN Redlist](https://www.iucnredlist.org/species/41698/45218636), okay to exclude for balance.\n", " - _felis lybica_ listed as least concern [IUCN](https://www.iucnredlist.org/species/131299383/154907281) (\"Afro-Asiatic Wildcat, one of the most common felid species\").\n", " - _felis catus_ is a domestic cat, not even considered by IUCN.\n", " - _alopochen aegyptiaca_ least concern, but decreasing [IUCN](https://www.iucnredlist.org/species/22679993/131910647).\n", " - _herpestes sanguineus_ least concern, stable [IUCN](https://www.iucnredlist.org/species/41606/45206143).\n", " - _bos taurus_ domestic cattle, represented in others probably not so important.\n", " - _proteles cristatus_ least concern, stable, but needs updating [IUCN](https://www.iucnredlist.org/species/18372/45195681).\n", "\n", "So, for the balanced set we'll remove those that aren't classified down to species, then maybe randomly select 22 per species or we could keep it consistent with the Ohio small animals and do 12. That'd give us either 660 or 372 image test set (30 species at 22 or 31 species at 12 images per species, respectively)." ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sns.histplot(df, y = 'class')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Filter Classifications not to Species" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "Index: 61040 entries, 0 to 63467\n", "Data columns (total 10 columns):\n", " # Column Non-Null Count Dtype \n", "--- ------ -------------- ----- \n", " 0 original_label 61040 non-null object\n", " 1 scientific_name 61040 non-null object\n", " 2 common_name 61040 non-null object\n", " 3 kingdom 61040 non-null object\n", " 4 phylum 61040 non-null object\n", " 5 class 61040 non-null object\n", " 6 order 61040 non-null object\n", " 7 family 61040 non-null object\n", " 8 genus 61040 non-null object\n", " 9 species 61040 non-null object\n", "dtypes: object(10)\n", "memory usage: 5.1+ MB\n" ] } ], "source": [ "df_filter = df.loc[df[\"species\"].notna()].copy()\n", "df_filter[taxa_cols].info(show_counts=True)" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "1" ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_filter[\"subspecies\"].nunique()" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "subspecies\n", "equus zebra hartmannae 11687\n", "Name: count, dtype: int64" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_filter[\"subspecies\"].value_counts()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can check if equus zebra is in here without subspecies designations." ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0" ] }, "execution_count": 20, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_filter.loc[(df_filter[\"species\"] == \"equus zebra\") & (df_filter[\"subspecies\"].isna())].shape[0]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Yes, this is the only way equus zebras are represented, so we can remove the subspecies." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Remove extra columns\n", "\n", "Only need `taxa_cols` (Linnean taxonomy + `original_label`, `scientific_name`, and `common_name`) and `id_cols`." ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "id_cols = ['dataset_name',\n", " 'url_gcp',\n", " 'url_aws',\n", " 'url_azure',\n", " 'image_id',\n", " 'sequence_id',\n", " 'location_id',\n", " 'frame_num']\n", "\n", "cols_to_keep = [col for col in list(df.columns) if (col in id_cols or col in taxa_cols)]" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "['dataset_name',\n", " 'url_gcp',\n", " 'url_aws',\n", " 'url_azure',\n", " 'image_id',\n", " 'sequence_id',\n", " 'location_id',\n", " 'frame_num',\n", " 'original_label',\n", " 'scientific_name',\n", " 'common_name',\n", " 'kingdom',\n", " 'phylum',\n", " 'class',\n", " 'order',\n", " 'family',\n", " 'genus',\n", " 'species']" ] }, "execution_count": 22, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cols_to_keep" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let's add a number of images column (by `scientific_name`)." ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
dataset_nameurl_gcpurl_awsurl_azureimage_idsequence_idlocation_idframe_numoriginal_labelscientific_name...familysubfamilytribegenusspeciessubspeciesvarietymulti_speciesnum_speciesnum_sp_images
0Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...felidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0450.0
1Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...felidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0450.0
2Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...felidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0450.0
3Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...felidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0450.0
4Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatus...felidaefelinaeNaNacinonyxacinonyx jubatusNaNNaNFalse1.0450.0
\n", "

5 rows × 35 columns

\n", "
" ], "text/plain": [ " dataset_name \\\n", "0 Desert Lion Conservation Camera Traps \n", "1 Desert Lion Conservation Camera Traps \n", "2 Desert Lion Conservation Camera Traps \n", "3 Desert Lion Conservation Camera Traps \n", "4 Desert Lion Conservation Camera Traps \n", "\n", " url_gcp \\\n", "0 https://storage.googleapis.com/public-datasets... \n", "1 https://storage.googleapis.com/public-datasets... \n", "2 https://storage.googleapis.com/public-datasets... \n", "3 https://storage.googleapis.com/public-datasets... \n", "4 https://storage.googleapis.com/public-datasets... \n", "\n", " url_aws \\\n", "0 http://us-west-2.opendata.source.coop.s3.amazo... \n", "1 http://us-west-2.opendata.source.coop.s3.amazo... \n", "2 http://us-west-2.opendata.source.coop.s3.amazo... \n", "3 http://us-west-2.opendata.source.coop.s3.amazo... \n", "4 http://us-west-2.opendata.source.coop.s3.amazo... \n", "\n", " url_azure \\\n", "0 https://lilawildlife.blob.core.windows.net/lil... \n", "1 https://lilawildlife.blob.core.windows.net/lil... \n", "2 https://lilawildlife.blob.core.windows.net/lil... \n", "3 https://lilawildlife.blob.core.windows.net/lil... \n", "4 https://lilawildlife.blob.core.windows.net/lil... \n", "\n", " image_id \\\n", "0 Desert Lion Conservation Camera Traps : acinon... \n", "1 Desert Lion Conservation Camera Traps : acinon... \n", "2 Desert Lion Conservation Camera Traps : acinon... \n", "3 Desert Lion Conservation Camera Traps : acinon... \n", "4 Desert Lion Conservation Camera Traps : acinon... \n", "\n", " sequence_id \\\n", "0 Desert Lion Conservation Camera Traps : unknown \n", "1 Desert Lion Conservation Camera Traps : unknown \n", "2 Desert Lion Conservation Camera Traps : unknown \n", "3 Desert Lion Conservation Camera Traps : unknown \n", "4 Desert Lion Conservation Camera Traps : unknown \n", "\n", " location_id frame_num \\\n", "0 Desert Lion Conservation Camera Traps : unknown -1 \n", "1 Desert Lion Conservation Camera Traps : unknown -1 \n", "2 Desert Lion Conservation Camera Traps : unknown -1 \n", "3 Desert Lion Conservation Camera Traps : unknown -1 \n", "4 Desert Lion Conservation Camera Traps : unknown -1 \n", "\n", " original_label scientific_name ... family subfamily tribe \\\n", "0 acinonyx jubatus acinonyx jubatus ... felidae felinae NaN \n", "1 acinonyx jubatus acinonyx jubatus ... felidae felinae NaN \n", "2 acinonyx jubatus acinonyx jubatus ... felidae felinae NaN \n", "3 acinonyx jubatus acinonyx jubatus ... felidae felinae NaN \n", "4 acinonyx jubatus acinonyx jubatus ... felidae felinae NaN \n", "\n", " genus species subspecies variety multi_species num_species \\\n", "0 acinonyx acinonyx jubatus NaN NaN False 1.0 \n", "1 acinonyx acinonyx jubatus NaN NaN False 1.0 \n", "2 acinonyx acinonyx jubatus NaN NaN False 1.0 \n", "3 acinonyx acinonyx jubatus NaN NaN False 1.0 \n", "4 acinonyx acinonyx jubatus NaN NaN False 1.0 \n", "\n", " num_sp_images \n", "0 450.0 \n", "1 450.0 \n", "2 450.0 \n", "3 450.0 \n", "4 450.0 \n", "\n", "[5 rows x 35 columns]" ] }, "execution_count": 23, "metadata": {}, "output_type": "execute_result" } ], "source": [ "for sci_name in list(df_filter[\"scientific_name\"].unique()):\n", " df_filter.loc[df_filter[\"scientific_name\"] == sci_name, \"num_sp_images\"] = df_filter.loc[df_filter[\"scientific_name\"] == sci_name].shape[0]\n", "\n", "df_filter.head()" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [], "source": [ "cols_to_keep.append(\"num_sp_images\")" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
dataset_nameurl_gcpurl_awsurl_azureimage_idsequence_idlocation_idframe_numoriginal_labelscientific_namecommon_namekingdomphylumclassorderfamilygenusspeciesnum_sp_images
0Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatuscheetahanimaliachordatamammaliacarnivorafelidaeacinonyxacinonyx jubatus450.0
1Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatuscheetahanimaliachordatamammaliacarnivorafelidaeacinonyxacinonyx jubatus450.0
2Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatuscheetahanimaliachordatamammaliacarnivorafelidaeacinonyxacinonyx jubatus450.0
3Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatuscheetahanimaliachordatamammaliacarnivorafelidaeacinonyxacinonyx jubatus450.0
4Desert Lion Conservation Camera Trapshttps://storage.googleapis.com/public-datasets...http://us-west-2.opendata.source.coop.s3.amazo...https://lilawildlife.blob.core.windows.net/lil...Desert Lion Conservation Camera Traps : acinon...Desert Lion Conservation Camera Traps : unknownDesert Lion Conservation Camera Traps : unknown-1acinonyx jubatusacinonyx jubatuscheetahanimaliachordatamammaliacarnivorafelidaeacinonyxacinonyx jubatus450.0
\n", "
" ], "text/plain": [ " dataset_name \\\n", "0 Desert Lion Conservation Camera Traps \n", "1 Desert Lion Conservation Camera Traps \n", "2 Desert Lion Conservation Camera Traps \n", "3 Desert Lion Conservation Camera Traps \n", "4 Desert Lion Conservation Camera Traps \n", "\n", " url_gcp \\\n", "0 https://storage.googleapis.com/public-datasets... \n", "1 https://storage.googleapis.com/public-datasets... \n", "2 https://storage.googleapis.com/public-datasets... \n", "3 https://storage.googleapis.com/public-datasets... \n", "4 https://storage.googleapis.com/public-datasets... \n", "\n", " url_aws \\\n", "0 http://us-west-2.opendata.source.coop.s3.amazo... \n", "1 http://us-west-2.opendata.source.coop.s3.amazo... \n", "2 http://us-west-2.opendata.source.coop.s3.amazo... \n", "3 http://us-west-2.opendata.source.coop.s3.amazo... \n", "4 http://us-west-2.opendata.source.coop.s3.amazo... \n", "\n", " url_azure \\\n", "0 https://lilawildlife.blob.core.windows.net/lil... \n", "1 https://lilawildlife.blob.core.windows.net/lil... \n", "2 https://lilawildlife.blob.core.windows.net/lil... \n", "3 https://lilawildlife.blob.core.windows.net/lil... \n", "4 https://lilawildlife.blob.core.windows.net/lil... \n", "\n", " image_id \\\n", "0 Desert Lion Conservation Camera Traps : acinon... \n", "1 Desert Lion Conservation Camera Traps : acinon... \n", "2 Desert Lion Conservation Camera Traps : acinon... \n", "3 Desert Lion Conservation Camera Traps : acinon... \n", "4 Desert Lion Conservation Camera Traps : acinon... \n", "\n", " sequence_id \\\n", "0 Desert Lion Conservation Camera Traps : unknown \n", "1 Desert Lion Conservation Camera Traps : unknown \n", "2 Desert Lion Conservation Camera Traps : unknown \n", "3 Desert Lion Conservation Camera Traps : unknown \n", "4 Desert Lion Conservation Camera Traps : unknown \n", "\n", " location_id frame_num \\\n", "0 Desert Lion Conservation Camera Traps : unknown -1 \n", "1 Desert Lion Conservation Camera Traps : unknown -1 \n", "2 Desert Lion Conservation Camera Traps : unknown -1 \n", "3 Desert Lion Conservation Camera Traps : unknown -1 \n", "4 Desert Lion Conservation Camera Traps : unknown -1 \n", "\n", " original_label scientific_name common_name kingdom phylum \\\n", "0 acinonyx jubatus acinonyx jubatus cheetah animalia chordata \n", "1 acinonyx jubatus acinonyx jubatus cheetah animalia chordata \n", "2 acinonyx jubatus acinonyx jubatus cheetah animalia chordata \n", "3 acinonyx jubatus acinonyx jubatus cheetah animalia chordata \n", "4 acinonyx jubatus acinonyx jubatus cheetah animalia chordata \n", "\n", " class order family genus species num_sp_images \n", "0 mammalia carnivora felidae acinonyx acinonyx jubatus 450.0 \n", "1 mammalia carnivora felidae acinonyx acinonyx jubatus 450.0 \n", "2 mammalia carnivora felidae acinonyx acinonyx jubatus 450.0 \n", "3 mammalia carnivora felidae acinonyx acinonyx jubatus 450.0 \n", "4 mammalia carnivora felidae acinonyx acinonyx jubatus 450.0 " ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_reduced = df_filter[cols_to_keep].copy()\n", "df_reduced.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Reduce to no more than 10K images per species" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "32854\n" ] } ], "source": [ "imgs_to_keep = list(df_reduced.loc[df_reduced[\"num_sp_images\"] <= 10000, \"image_id\"])\n", "print(len(imgs_to_keep))" ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "2" ] }, "execution_count": 27, "metadata": {}, "output_type": "execute_result" } ], "source": [ "high_num_classes = list(df_reduced.loc[~df_reduced[\"image_id\"].isin(imgs_to_keep), \"scientific_name\"].unique())\n", "len(high_num_classes)" ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "52854" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" } ], "source": [ "for sci_name in high_num_classes:\n", " sample_set = list(df_reduced.loc[df_reduced[\"scientific_name\"] == sci_name].sample(10000, random_state = 614)[\"image_id\"])\n", " imgs_to_keep = imgs_to_keep + sample_set\n", "\n", "len(imgs_to_keep)" ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [], "source": [ "df_reduced_upper_bound = df_reduced.loc[df_reduced[\"image_id\"].isin(imgs_to_keep)].copy()\n", "df_reduced_upper_bound.to_csv(\"../data/potential-test-sets/filtered/desert-lion-upper-bound.csv\", index = False)" ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [], "source": [ "df_reduced_upper_lower = df_reduced_upper_bound.loc[df_reduced_upper_bound[\"num_sp_images\"] >= 10].copy()\n", "df_reduced_upper_lower.to_csv(\"../data/potential-test-sets/filtered/desert-lion-upper-lower-bound.csv\", index = False)" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sns.histplot(df_reduced_upper_lower, y = \"scientific_name\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Randomly sample to balanced set (12 images per species)\n" ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "372" ] }, "execution_count": 31, "metadata": {}, "output_type": "execute_result" } ], "source": [ "balanced_set = []\n", "for sci_name in list(df_reduced[\"scientific_name\"].unique()):\n", " temp = df_reduced.loc[df_reduced[\"scientific_name\"] == sci_name].copy()\n", " if temp.shape[0] < 12:\n", " continue\n", " sample_set = list(temp.sample(12, random_state = 614)[\"image_id\"])\n", " balanced_set = balanced_set + sample_set\n", "\n", "len(balanced_set)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Filter to just balanced set and drop the number of species column since it's been balanced to 12 each." ] }, { "cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [], "source": [ "df_balanced = df_reduced.loc[df_reduced[\"image_id\"].isin(balanced_set)].copy()\n", "df_balanced.drop(columns = [\"num_sp_images\"], inplace = True)\n", "df_balanced.to_csv(\"../data/potential-test-sets/filtered/desert-lion-balanced.csv\", index = False)" ] }, { "cell_type": "code", "execution_count": 33, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "dataset_name 1\n", "url_gcp 372\n", "url_aws 372\n", "url_azure 372\n", "image_id 372\n", "sequence_id 1\n", "location_id 1\n", "frame_num 1\n", "original_label 31\n", "scientific_name 31\n", "common_name 31\n", "kingdom 1\n", "phylum 1\n", "class 2\n", "order 12\n", "family 17\n", "genus 28\n", "species 31\n", "dtype: int64" ] }, "execution_count": 33, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_balanced.nunique()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "jupytext": { "formats": "ipynb,py:percent" }, "kernelspec": { "display_name": "data-dev", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.12.3" } }, "nbformat": 4, "nbformat_minor": 2 }