Streamlit_Create_prediction_app.md

Template request | Bug report | Generate Data Product

Tags: #streamlit #app #ml #ai #operations #plotly

Author: Gagan Bhatia

Description: This notebook provides a step-by-step guide to creating a Streamlit app that can make predictions based on user input.

Input

Import library

from naas_drivers import streamlit

Model

Create the Python file necessary to deploy Streamlit app.

%%writefile streamlit_app.py

from naas_drivers import streamlit, plotly, yahoofinance, prediction
import streamlit as st

TICKER = "TSLA"
date_from = -100 # 1OO days max to feed the naas_driver for prediction
date_to = "today"
DATA_POINT = 20

df_yahoo = yahoofinance.get(tickers=TICKER,
                            date_from=date_from,
                            date_to=date_to).dropna().reset_index(drop=True)

df_predict = prediction.get(dataset=df_yahoo,
                            date_column='Date',
                            column="Close",
                            data_points=DATA_POINT,
                            prediction_type="all").sort_values("Date", ascending=False).reset_index(drop=True)

fig = plotly.linechart(df_predict,
                       x="Date",
                       y=["Close", "ARIMA", "SVR", "LINEAR", "COMPOUND"],
                       showlegend=True,
                       title=f"{TICKER} predictions as of today, for next {str(DATA_POINT)} days.")



st.write("# Prediction for {}".format(TICKER))
st.plotly_chart(fig, width=1200)

Output

Deploy the app from Python file and serve the URL where the app is exposed.

streamlit.add("streamlit_app.py", port=9999, debug=True)