Weather

Powerball draws 3 days per week so Lotto Bot will use the datetime library to find the next drawing.

import datetime
from datetime import timedelta
from datetime import datetime
import pytz

utc=pytz.utc                                                                                                                                                                                         
nextdraw_date = datetime.today()
 
while nextdraw_date.weekday() != 0:
    if nextdraw_date.weekday() == 2:
        break
    if nextdraw_date.weekday() == 5:
        break
    else:
        nextdraw_date += timedelta(1)

nextdraw_date = nextdraw_date.replace(hour=22, minute=0)
nextdraw_enddate = nextdraw_date.replace(hour=23, minute=0)
nextdraw_date_utcepoch = int(nextdraw_date.astimezone(utc).timestamp())

Lotto Bot is using the request library to talk to a weather API to get the forecast for the drawing location. Powerball is drawn at 250 Marriott Dr | Tallahassee, FL so Lotto Bot requests weather data for that location on the draw date.

import requests
import json

response = requests.request("GET", "[some weather API]")
forecast_data = response.json()

try: 
    Prediction.objects.update_or_create(
        draw_date = str(nextdraw_date.strftime("%Y-%m-%d")),
        draw_night = str(nextdraw_date.weekday()),
        Barometer = forecast_data['Barometer'],
        Dewpoint = forecast_data['Dewpoint'],
        Heat_Index = forecast_data['Heat_Index'],
        Hygrometer = forecast_data['Hygrometer'],
        Thermometer = forecast_data['Thermometer'],
        Percipitation = forecast_data['Percipitation'],
        WindDirection = forecast_data['WindDirection'],
        CloudCover = forecast_data['CloudCover'],
        WindSpeed = forecast_data['WindSpeed'],
        MoonPhase = forecast_data['MoonPhase'],
)

Here Lotto Bot must use some predictive data analysis to find similarities in the forecasted weather patterns and the historic data.

import numpy as np
import pandas as pd

import sklearn
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn import linear_model
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.impute import SimpleImputer

# put rows in a dataframe                                                                                                                                                                            
df_w = pd.DataFrame([Lottery Drawing History Data])

for column in df_w:
    df_w[column] = pd.to_numeric(df_w[column], errors='coerce')

df_w.head()

# Split the data into independent and dependent variables
X = df_w.iloc[:,6:16].values
y = df_w.iloc[:,0:6].values

# Creating the Training and Test set from data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2)#, random_state = 69)

# Feature Scaling
scaler = StandardScaler()
X_fit_train = scaler.fit_transform(X_train)
X_fit_test = scaler.transform(X_test)
X_fit = scaler.fit_transform(X)

# Fitting Random Forest Classification to the Training set
RFclassifier = RandomForestClassifier(n_estimators=24, criterion="entropy", random_state=12)
RFclassifier.fit(X_fit_train, y_train)

# Compare the Test set results for Random Forest Classification
y_pred = RFclassifier.predict(X_fit_test)
accuracyScore = 0
totalMatchScore = 0

for index, pred in enumerate(y_pred):
    accuracy = accuracy_score(y_test[index], pred)
    accuracyScore += accuracy

    comparePOS = list(set(pred[:5]) & set(y_test[index][:5]))
    comparePWB = list(set(pred[5:]) & set(y_test[index][5:]))