{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"collapsed":true},"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 5GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# **1. Generar diccionario de las fotos**","execution_count":null},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"import csv\ntrain = {}\nfor i in range(5):\n    train[i] = []\n    \nprueba = []\nwith open('/kaggle/input/aptos2019-blindness-detection/train.csv', mode='r') as csv_file:\n    csv_reader = csv.DictReader(csv_file)\n    line_count = 0\n    for row in csv_reader:\n        if line_count == 0:\n            print(f'Column names are {\", \".join(row)}')\n        #print(row['diagnosis'],row['id_code'])\n        #prueba.append(int(row['diagnosis']) )\n        train[ int(row['diagnosis']) ].append(row['id_code'])\n        line_count += 1\n    \n    print(f'Processed {line_count} lines.')\n\n#print(train)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"test = {}\nfor i in range(5):\n    test[i] = []\n    \nprueba = []\nwith open('/kaggle/input/aptos2019-blindness-detection/sample_submission.csv', mode='r') as csv_file:\n    csv_reader = csv.DictReader(csv_file)\n    line_count = 0\n    for row in csv_reader:\n        if line_count == 0:\n            print(f'Column names are {\", \".join(row)}')\n        test[ int(row['diagnosis']) ].append(row['id_code'])\n        #print(int(row['diagnosis']) ,row['id_code'])\n        line_count += 1\n    \n    print(f'Processed {line_count} lines.')\n#print(test)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# **2. Genración de las graficas**","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"import matplotlib.pyplot as plt\nx_units = [1, 2, 3, 4, 5]\ny_units = []\nfor i in train:\n    y_units.append(len(train[i]))\n\ntick_label = ['0', '1', '2', '3', '4']\n\nplt.bar(x_units, y_units, tick_label=tick_label,\n        width=0.8)\n\nplt.xlabel('Clasificación')\nplt.ylabel('Cantidad de imágenes')\nplt.title('Distribución train')\n\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Como se puede observar la cantidad de fotos no se encuentra balanceada entre las diferentes clases.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"x_units = [1, 2, 3, 4, 5]\ny_units = []\nfor i in test:\n    y_units.append(len(test[i]))\n\ntick_label = ['0', '1', '2', '3', '4']\n\nplt.bar(x_units, y_units, tick_label=tick_label,\n        width=0.8)\n\nplt.xlabel('Clasificación')\nplt.ylabel('Cantidad de imágenes')\nplt.title('Distribución test')\n\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Como se puede observar el set de test, solo tiene fotos de una categoria. ","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"# **3. Balanceo de la data**","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"print(len(train[0]))\nprint(len(train[1]), len(train[0])/len(train[1]))\nprint(len(train[2]), len(train[0])/len(train[2]))\nprint(len(train[3]), len(train[0])/len(train[3]))\nprint(len(train[4]), len(train[0])/len(train[4]))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Cantidades por categoria, y factor a utilizar para balancear.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#balanceo\nimport itertools\nvalues = [1,5,2,9,6]\nfor i in range(len(values)):\n    print(values[i])\n    temp = list(itertools.repeat(train[i], values[i]))\n    resul = []\n    for j in temp:\n        resul = resul + j\n    train[i] = resul","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"x_units = [1, 2, 3, 4, 5]\ny_units = []\nfor i in train:\n    y_units.append(len(train[i]))\n\ntick_label = ['0', '1', '2', '3', '4']\n\nplt.bar(x_units, y_units, tick_label=tick_label,\n        width=0.8)\n\nplt.xlabel('Clasificación')\nplt.ylabel('Cantidad de imágenes')\nplt.title('Distribución train')\n\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Como podemos ver el dataset , ya se encuentra balanceado\n","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"Cantidades ya balanceadas","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"print(len(train[0]))\nprint(len(train[1]))\nprint(len(train[2]))\nprint(len(train[3]))\nprint(len(train[4]))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# **4. Comparción de tamaños**","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"import PIL\nsizes = {}\nfor i in range(5):\n    for j in train[i]:    \n        image = PIL.Image.open(\"/kaggle/input/aptos2019-blindness-detection/train_images/\"+j+\".png\")\n        width, height = image.size\n        #print(width, height)\n        if (width, height) in sizes:\n            sizes[width, height].append(j)\n        else:\n            sizes[width, height] = []\n            sizes[width, height].append(j)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"x_units = []\nfor i in range(len(sizes)):\n    x_units.append(i)\n#print(x_units)\ny_units = []\nfor i, key in enumerate(sizes):\n    print(i, key)\n    y_units.append(len(sizes[key]))\n\ntick_label = []\n\nfor i in range(len(sizes)):\n    tick_label.append(i)\n    \n#for key in sizes:\n#    tick_label.append(key)\n    \nplt.bar(x_units, y_units, tick_label=tick_label,\n         width=0.7)\n\nplt.xlabel('No. de resolucion')\nplt.ylabel('Cantidad de imágenes')\nplt.title('Distribución train')\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Como podemos observar la gran mayoría de las imagenes cuentan con altas resoluciones. Siendo 2416x1736, la más repetida. ","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"import PIL\nsizes = {}\nfor i in range(5):\n    for j in test[i]:    \n        image = PIL.Image.open(\"/kaggle/input/aptos2019-blindness-detection/test_images/\"+j+\".png\")\n        width, height = image.size\n        #print(width, height)\n        if (width, height) in sizes:\n            sizes[width, height].append(j)\n        else:\n            sizes[width, height] = []\n            sizes[width, height].append(j)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"x_units = []\nfor i in range(len(sizes)):\n    x_units.append(i)\n#print(x_units)\ny_units = []\nfor i, key in enumerate(sizes):\n    print(i, key)\n    y_units.append(len(sizes[key]))\n\ntick_label = []\n\nfor i in range(len(sizes)):\n    tick_label.append(i)\n    \n#for key in sizes:\n#    tick_label.append(key)\n    \nplt.bar(x_units, y_units, tick_label=tick_label,\n         width=0.7)\n\nplt.xlabel('Clasificación')\nplt.ylabel('Cantidad de imagenes')\nplt.title('Distribución test')\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"En el caso del testing, la mayoria de las imagenes cuentan con un tamaño medio de 640 x 480.\n","execution_count":null}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}