run_ttest.py

# Running ttest -->
def run_ttest():
        #Get All relevant constraints -->
        housing_data = convert_housing_data_to_quarters()
        startQ = get_recession_start()
        bottomQ = get_recession_bottom()
        endQ = get_recession_end()

        #Recession Timeframe -->
        housing_data['Delta'] = housing_data[bottomQ]- housing_data[startQ]
                
        #Regions with Universities in them -->
        uni_regions = get_list_of_university_towns()
        uni_regions['Uni_regions'] = True        
        
        #merge housing data with uni_town df -->
        data2 = pd.merge(housing_data,uni_regions,how='right', left_index=True,right_on=['State', 'RegionName'])
        uni_towns = data2[data2['Uni_regions']==True]['Delta']
            
        #merge housing data with non-uni_town df -->
        data3 = pd.merge(data,uni_regions,how='left',left_index=True,right_on=['State','RegionName'])
        data3['Uni_regions'] = data3['Uni_regions'].replace({np.NaN: False})
        non_uni_towns = data3[data3['Uni_regions'] == False]['Delta']

        #Run ttest on both dataframes -->
        st,p = stats.ttest_ind(uni_towns,non_uni_towns,nan_policy='omit')
        
        #get different p-values (bar at 0.01) -->
        different = False 
        if p < 0.01:
            different = True     # p<0.01 means lower than 1% of the thing happening by chance.
         
        #Which place is better -->
        better = ""
        if uni_towns.mean() > non_uni_towns.mean():
            better = "university town"
        else:
            better = "non-university towns"
        
        return (different,p,better)
            
            
run_ttest()