corona_features.ipynb

{
    "cells": [
        {
            "cell_type": "code",
            "execution_count": 1,
            "metadata": {},
            "outputs": [],
            "source": "# The code was removed by Watson Studio for sharing."
        },
        {
            "cell_type": "code",
            "execution_count": 2,
            "metadata": {},
            "outputs": [],
            "source": "import types\nimport pandas as pd\nfrom botocore.client import Config\nimport ibm_boto3\n\ndef __iter__(self): return 0\n\n\nclient = ibm_boto3.client(service_name='s3',\n    ibm_api_key_id=credentials_1['IBM_API_KEY_ID'],\n    ibm_auth_endpoint=credentials_1['IBM_AUTH_ENDPOINT'],\n    config=Config(signature_version='oauth'),\n    endpoint_url=credentials_1['ENDPOINT'])\nbucket=credentials_1['BUCKET']"
        },
        {
            "cell_type": "code",
            "execution_count": 3,
            "metadata": {},
            "outputs": [],
            "source": "client.download_file(Bucket=bucket,Key='cases.csv',Filename='cases.csv')"
        },
        {
            "cell_type": "code",
            "execution_count": 4,
            "metadata": {},
            "outputs": [],
            "source": "df_raw = pd.read_csv('cases.csv')"
        },
        {
            "cell_type": "code",
            "execution_count": 5,
            "metadata": {},
            "outputs": [
                {
                    "data": {
                        "text/plain": "['Austria',\n 'Belgium',\n 'Canada',\n 'Czechia',\n 'Denmark',\n 'Dominican_Republic',\n 'Ecuador',\n 'France',\n 'Germany',\n 'Hungary',\n 'Iran',\n 'Ireland',\n 'Israel',\n 'Italy',\n 'Netherlands',\n 'Norway',\n 'Portugal',\n 'Romania',\n 'Spain',\n 'Sweden',\n 'Switzerland',\n 'Turkey',\n 'United_Kingdom',\n 'United_States_of_America']"
                    },
                    "execution_count": 5,
                    "metadata": {},
                    "output_type": "execute_result"
                }
            ],
            "source": "window_size = 10\n\ndf = df_raw\n\n#df = df[df['countriesAndTerritories']=='Italy']\n\ndf = df.sort_values(['year', 'month' ,'day'], ascending=[1, 1, 1])\n\ndf['ts'] = pd.to_datetime(df[['year', 'month' ,'day']])\n\ndf['increase_cases'] = df.cases - df.cases.shift(1)\n\ndf['increase_ratio'] = df.cases / df.cases.shift(1)\n\ndf['cases_estimated'] = df.deaths*100\n\ndf['total_cases_estimated'] = df['cases_estimated'].cumsum()\n\ndf['percentage_infected'] = 100/df[\"2018\"]*df['total_cases_estimated']\n\ndf['percentage_died'] = 100/df[\"2018\"]*df['deaths']\n\n\nfor i in range(1,window_size+1):\n    df['cases_'+str(i)+'_days_before'] = df.cases.shift(i)\n    \n    \nfor i in range(1,window_size+1):\n    df['percentage_died_'+str(i)+'_days_before'] = df.percentage_died.shift(i)\n    \n        \nfor i in range(1,window_size+1):\n    df['deaths_'+str(i)+'_days_before'] = df.deaths.shift(i)\n\naverage = 0\nfor i in range(1,window_size+1):\n    average = average + df.cases.shift(i)\n    \ndf['cases_'+str(window_size)+'_day_average'] = average/window_size\n\naverage = 0\nfor i in range(1,window_size+1):\n    average = average + df.percentage_died.shift(i)\n    \ndf['percentage_died_'+str(window_size)+'_day_average'] = average/window_size\n\naverage = 0\nfor i in range(1,window_size+1):\n    average = average + df.deaths.shift(i)\n    \ndf['deaths_'+str(window_size)+'_day_average'] = average/window_size\n\n#for i in range(1,window_size+1):\n#    df = df[df['deaths_'+str(i)+'_days_before']>0]\n\n#df = df[df['Cases']>100]\n\n\n#df = df[df['ts']>'2020-04-01']\n\ngroup_by_deaths = df.groupby(['countriesAndTerritories']).sum()['percentage_died']\nbad_countries = list(group_by_deaths[group_by_deaths>0.001].index)\n\n\n\n\n#df = df[df['countriesAndTerritories'].isin(bad_countries)]\n#df\nbad_countries\n\n\n#Germany                      1607\n#Italy                       15253\n#Netherlands                  1538\n#Spain                       11570\n#United_Kingdom               4972\n#United_States_of_America    10973"
        },
        {
            "cell_type": "code",
            "execution_count": 6,
            "metadata": {},
            "outputs": [],
            "source": "df.to_csv('cases_features.csv')"
        },
        {
            "cell_type": "code",
            "execution_count": 7,
            "metadata": {},
            "outputs": [],
            "source": "client.upload_file(Filename='cases_features.csv',Bucket=bucket,Key='cases_features.csv')"
        }
    ],
    "metadata": {
        "kernelspec": {
            "display_name": "Python 3.6",
            "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": "3.6.9"
        }
    },
    "nbformat": 4,
    "nbformat_minor": 4
}

Hi Romeo,
I was wondering in section 5: while first sorting your csv at date and then using the shift function to calculate the increase_cases column, aren't you comparing cases from different countries instead of different days? As I'm now reading it, I would assume we want to compare the cases for the same country.