Crime-Prediction

Crime Prediction

-crimepred.ipynb can be opened on jupyter notebook
-crimepred.py can be executed directly a executable file
-Since data can be fetched from website using API (which is implemented in the provided source code). For that user needs to install sodapy library using following command:
pip install sodapy
-Also, it can be downloaded from the following link:
https://data.sfgov.org/Public-Safety/-Change-Notice-Police-Department-Incidents/tmnf-yvry

Libraries Used:
-Python2.7
-pandas
-Numpy
-sklearn
-sodapy

Program can be executed in sequence in jupyter
Following steps to execute program:
-Execute everything till you reach to “Method for Classification”
-From here there are methods developed for the ease of user to train classification model and perfirm prediction on testing data
-predictCrimeCategory() can be executed to perform classification for specific district
```
  Parameters:
      -district name
      -Classifier
```
-classifyAllDistricts() can be executed to perform classification for all the districts in San Francisco
```
  Parameters:
      -Classifier
      -path where output needs to be stored
```
-ensamble_classifiers() can be executed for city as well as district and it generates all 3 classifier predictions and from that performs ensembling in order to generate better prediction
```
  Parameters:
      -district name
      -city - if true classification will be performed on city level else on district level
```