python-time-series-forecasting/main.py

import matplotlib.pyplot as plt
import pandas as pd
from statsmodels.tsa.holtwinters import ExponentialSmoothing
import requests
import json

# ts = pd.read_csv("d:\\files\\Features count.csv")
# ts.plot()
# plt.grid()
# plt.show()

# moving_avg = ts.rolling(4).mean()
# ts.plot()
# plt.plot(moving_avg, color='green', label='Moving average')
# plt.legend()
# plt.grid()

# model = ExponentialSmoothing(ts, trend="additive", seasonal="additive", seasonal_periods=5)
# fit1 = model.fit()
# pred1 = fit1.forecast(5)

# plt.plot(ts, label='Features count')
# plt.plot(pred1, color='red', label='Holt-Winters forecast')
# plt.legend()
# plt.grid()

ts = pd.read_csv("d:\\files\\Testing time.csv")
model = ExponentialSmoothing(ts, trend="additive", seasonal="additive", seasonal_periods=7)
fit1 = model.fit()
pred1 = fit1.forecast(10)

plt.plot(ts, label='Testing time')
plt.plot(pred1, color='red', label='Holt-Winters forecast')
plt.legend()
plt.grid()
plt.show()

# Использовать внешний сервис прогнозирования
time_series = {
    "originalTimeSeries": {
        "values": [
            {
                "date": "2023-04-16T06:29:50.668476",
                "value": 1
            },
            {
                "date": "2023-04-17T06:29:50.668476",
                "value": 2
            },
            {
                "date": "2023-04-18T06:29:50.668476",
                "value": 3
            },
            {
                "date": "2023-04-19T06:29:50.668476",
                "value": 4
            },
            {
                "date": "2023-04-20T06:29:50.668476",
                "value": 5
            },
            {
                "date": "2023-04-21T06:29:50.668476",
                "value": 6
            },
            {
                "date": "2023-04-22T06:29:50.668476",
                "value": 5
            },
            {
                "date": "2023-04-23T06:29:50.668476",
                "value": 4
            },
            {
                "date": "2023-04-24T06:29:50.668476",
                "value": 3
            },
            {
                "date": "2023-04-25T06:29:50.668476",
                "value": 2
            },
            {
                "date": "2023-04-26T06:29:50.668476",
                "value": 1
            },
            {
                "date": "2023-04-27T06:29:50.668476",
                "value": 2
            }
        ],
        "name": 'test'
    },
    "countForecast": 5
}

source = []
for v in time_series['originalTimeSeries']['values']:
    source.append(v['value'])

url = "http://time-series.athene.tech/api/1.0/getForecast"
headers = {'Content-type': 'application/json'}
r = requests.post(url, data=json.dumps(time_series), headers=headers)

forecast = [item['value'] for item in r.json()['timeSeries']['values']]
plt.plot(range(1, 1+len(source)), source)
plt.plot(range(len(source), len(source)+len(forecast)), forecast)
plt.grid()
plt.show()