Add dockerfile and python script

Dockerfile

@@ -0,0 +1,21 @@
+FROM ubuntu:latest
+
+ENV KAGGLE_USERNAME="jaszwajcar"
+ENV KAGGLE_KEY="key"
+
+RUN apt update && apt install -y figlet
+RUN apt install unzip -y
+RUN apt install python3-pip -y
+
+RUN pip3 install kaggle
+RUN pip3 install pandas
+RUN pip3 install unzip
+RUN pip3 install scikit-learn
+RUN pip3 install seaborn
+
+WORKDIR /app
+
+COPY ./download_dataset.sh ./
+COPY ./stats.sh ./
+COPY ./data.py ./
+

data.py

@@ -0,0 +1,24 @@
+import pandas as pd
+from sklearn.model_selection import train_test_split
+
+baltimore=pd.read_csv('BPD_Part_1_Victim_Based_Crime_Data.csv')
+baltimore["Weapon"].fillna("None", inplace=True)
+baltimore.dropna(inplace=True)
+
+baltimore['Post'] = baltimore['Post'] /baltimore['Post'].abs().max()
+baltimore['Location']=baltimore['Location'].str.lower()
+baltimore['Description']=baltimore['Description'].str.lower()
+baltimore['Weapon']=baltimore['Weapon'].str.lower()
+baltimore['Premise']=baltimore['Premise'].str.lower()
+baltimore['District']=baltimore['District'].str.lower()
+baltimore['CrimeCode']=baltimore['CrimeCode'].str.lower()
+baltimore['Neighborhood']=baltimore['Neighborhood'].str.lower()
+baltimore['Inside/Outside']=baltimore['Inside/Outside'].str.lower()
+
+
+baltimore_train, baltimore_test = train_test_split(baltimore, test_size=0.1, random_state=1)
+baltimore_train, baltimore_dev= train_test_split(baltimore_train, test_size=0.25, random_state=1)
+
+baltimore_test.to_csv("baltimore_test.csv", encoding="utf-8", index=False)
+baltimore_dev.to_csv("baltimore_dev.csv", encoding="utf-8", index=False)
+baltimore_train.to_csv("baltimore_train.csv", encoding="utf-8", index=False)

download_dataset.sh

@@ -1,6 +1,4 @@
-kaggle datasets download -d sohier/crime-in-baltimore
+#!/bin/bash
-unzip crime-in-baltimore.zip
 
-head -n 8000 BPD_Part_1_Victim_Based_Crime_Data.csv > baltimore_train.csv
+kaggle datasets download -d sohier/crime-in-baltimore
-tail -n 2000 BPD_Part_1_Victim_Based_Crime_Data.csv > baltimore_test.csv
+unzip crime-in-baltimore.zip
-head -n 2000 baltimore_train.csv > baltimore_dev.csv