Stock technical analysis dashboard with Streamlit, yfinance and pandas-ta

This is a one-file Streamlit app for eyeballing a stock: pick a ticker and a date range in the sidebar, and it charts moving averages, Bollinger Bands, RSI and price-volume. Streamlit turns a top-to-bottom Python script into a web page and reruns it whenever an input changes, so there is no front-end code. yfinance pulls daily prices from Yahoo Finance, and pandas-ta adds indicators as new DataFrame columns.
The first version of this app (2023) stopped working on Streamlit Cloud, and it no longer runs anywhere: it called yf.pdr_override() and st.beta_columns(), both since removed, and it charted an adj_close column that yfinance never returned. The version below was rewritten and tested against Streamlit 1.67, yfinance 1.8, pandas 3.1 and pandas-ta-classic 0.9.
Install
Section titled “Install”pip install streamlit yfinance pandas-ta-classicThe app
Section titled “The app”Save this as app.py:
import datetime
import pandas_ta_classic as ta # registers the DataFrame.ta accessorimport streamlit as stimport yfinance as yf
st.title("Technical analysis dashboard")
st.sidebar.header("Inputs")ticker = st.sidebar.text_input("Ticker", "RY.TO")start = st.sidebar.date_input("Start date", datetime.date(2019, 1, 1))end = st.sidebar.date_input("End date", datetime.date.today())length = st.sidebar.number_input("Indicator length", min_value=2, max_value=200, value=20)
@st.cache_data(ttl=3600)def load(ticker: str, start: datetime.date, end: datetime.date): # multi_level_index=False keeps flat column names (Close, Volume, ...) return yf.download(ticker, start=start, end=end, progress=False, multi_level_index=False)
data = load(ticker, start, end)if data.empty: st.error(f"No data returned for {ticker}") st.stop()
data = data.copy()data.ta.sma(length=length, append=True)data.ta.ema(length=length, append=True)data.ta.bbands(length=length, append=True)data.ta.rsi(length=length, append=True)data.ta.pvol(append=True)
st.header(f"{ticker}: close vs. SMA and EMA")st.line_chart(data[["Close", f"SMA_{length}", f"EMA_{length}"]])
st.header("Bollinger Bands")st.line_chart(data[["Close", f"BBL_{length}_2.0", f"BBM_{length}_2.0", f"BBU_{length}_2.0"]])
st.header("RSI")st.line_chart(data[f"RSI_{length}"])
st.header("Price-volume")col1, col2 = st.columns(2)col1.line_chart(data["PVOL"])col2.bar_chart(data["Volume"])
with st.expander("Raw data"): st.dataframe(data)Run it with:
streamlit run app.pyYahoo tickers for Canadian listings need an exchange suffix: .TO for the TSX, .V for the TSX Venture and .CN for the CSE. The default here is RBC (RY.TO).
How the pandas-ta calls work
Section titled “How the pandas-ta calls work”Importing the library registers a .ta accessor on every DataFrame. Each indicator method finds the open/high/low/close/volume columns by name (case-insensitive), and append=True adds the result to the same frame instead of returning a new one. The new columns are named after the indicator and its parameters:
| Call | Columns added |
|---|---|
ta.sma(length=20) | SMA_20 |
ta.ema(length=20) | EMA_20 |
ta.bbands(length=20) | BBL_20_2.0, BBM_20_2.0, BBU_20_2.0, plus bandwidth BBB_20_2.0 and %B BBP_20_2.0 |
ta.rsi(length=20) | RSI_20 |
ta.pvol() | PVOL (close × volume) |
That naming is why the chart code builds column names with f-strings. If you change a default such as the Bollinger standard deviation (std=2.5), the suffix changes too (BBL_20_2.5).
Gotchas
Section titled “Gotchas”- yfinance column layout. Current yfinance releases return MultiIndex columns like
('Close', 'RY.TO')fromyf.download, even for one ticker. Passmulti_level_index=Falseor everydata["Close"]lookup returns a DataFrame instead of a Series. - Adjusted prices.
auto_adjust=Trueis now the default inyf.download, soCloseis already split- and dividend-adjusted and there is noAdj Closecolumn. Passauto_adjust=Falseif you want both. - Caching. Without
@st.cache_data, every widget change re-downloads the data. Yahoo rate-limits aggressive clients, and that is a common reason these apps break on shared hosting. - Indicator length vs. data range. A 200-day SMA on six months of data is all
NaN. Streamlit charts the gaps as blanks rather than raising an error.
Further reading
Section titled “Further reading”© 2026 David Li · Home · Blog · RSS ·Classic site · Built with Astro, Starlight and three.js · Logos:Font Awesome Free (CC BY 4.0)