Custom Charts

Often you want to visualize some metrics to are specific to your strategy and approach to developing algo-trading solutions. This example shows how to approach such a use-case. It is more meant as inspiration than a ready-to-use solution, but it should give you a good idea of how to use the roboquant framework to create your own custom charts.

It shows how to use the IndicatorMetric and SignalRatingMetric to track indicators and signals on a chart. It uses the YahooFeed to get the data for TSLA and plots the price, Bollinger Bands, RSI, and buy/sell signals on a chart.

We start with importing the required packages and modules.

import roboquant as rq
import matplotlib.pyplot as plt

from roboquant.util.metrics import IndicatorMetric, SignalRatingMetric
from roboquant.util.indicators import BBANDS, RSI

# set a default style
rq.set_light_style()

Like always, we need to create a feed to get the data for the chart. In this case we use the YahooFeed to get the data for Tesla for the year 2025.

feed = rq.feeds.YahooFeed("TSLA", start_date="2025-01-01", end_date="2026-01-01")
asset = feed.get_asset("TSLA")

Custom Metrics¶

We define the custom metrics we want to use. In both cases we subclass the IndicatorMetric and implement the _calc method. This method is called for each new bar in the feed, and is passed a BarSeries object that contains the last n bars, where n is the timeperiod of the indicator.

The indicators we use are the Bollinger Bands and the Relative Strength Index (RSI). The Bollinger Bands are calculated using the BBANDS function, which returns the upper, middle and lower bands. The RSI is calculated using the RSI function, which returns a single value.

Both functions are provided by the roboquant.util.indicators module, which is a wrapper around the TA-Lib library. The TA-Lib library is a popular library for technical analysis, and provides a wide range of indicators and functions.

class RSIMetric(IndicatorMetric):

    def _calc(self, buffer):
        return {"rsi": RSI(buffer.close(), self.timeperiod-1)}


class BBandsMetric(IndicatorMetric):

    def _calc(self, buffer):
        upper, _, lower = BBANDS(buffer.close(), timeperiod=self.timeperiod - 1)
        return {"lower": lower, "upper": upper}

Layout¶

We now define the figure and axes for the chart. We use a 3 row layout, with the first row for the price and Bollinger Bands, the second row for the RSI, and the third row for the buy/sell signals. The first row is 4 times the height of the other rows, to give more space for the price chart.

fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, sharex=True, height_ratios=[4,1,1])

Price Chart and Bollinger Bands¶

Lets plot the price and Bollinger Bands on the first axis. We use the track method of the feed to track the Bollinger Bands metric, which returns a TimeSeries DataFrame with the lower and upper bands.

# plot the prices first, without the volume.
feed.plot(asset, ax=ax1 , plot_volume=False)

metric = BBandsMetric(asset, timeperiod=10)
bbands = feed.track(metric)

ax1.fill_between(bbands.index, bbands["lower"], bbands["upper"], alpha=0.4, color="grey")
ax1.set_title(asset.symbol)

RSI Chart¶

Now we plot the RSI on the second axis. We use again the track method of the feed to track the RSI metric, which returns a TimeSeries DataFrame with the RSI values. We also add horizontal lines at 70 and 30 to indicate overbought and oversold conditions.

rsi = feed.track(RSIMetric(asset, timeperiod=10))
rsi.plot(ax=ax2)

ax2.axhline(70, color="red", linestyle="--")
ax2.axhline(30, color="green", linestyle="--")
ax2.set_yticks([])
ax2.grid(axis='y')
ax2.legend()

Buy/Sell Signals Chart¶

Finally we plot the buy/sell signals on the third axis. We use the SignalRatingMetric to track the signals generated by an EMA Crossover strategy. The metric returns a TimeSeries DataFrame with the ratings for each signal, which we use to plot the buy and sell signals as green and red bars respectively.

strategy = rq.strategies.EMACrossover(2, 5)
metric = SignalRatingMetric(asset, strategy = strategy)

ratings = feed.track(metric)
buy_ratings = ratings[ratings["rating/tsla"] > 0]
sell_ratings = ratings[ratings["rating/tsla"] < 0]

ax3.bar(buy_ratings.index, buy_ratings["rating/tsla"], color="green", label="buy")
ax3.bar(sell_ratings.index, sell_ratings["rating/tsla"], color="red", label="sell")
ax3.legend()

Plot¶

Now we can adjust the layout of the figure to make sure the subplots don’t overlap and display the chart.

fig.tight_layout(h_pad=0)
fig

Next Steps¶

Based on this example you can create your own custom charts to visualize any metrics you want.

You can also use the plot method of the feed to plot multiple assets on the same chart, or use the plot method of the metric to plot the metric values directly.

You can then combine these plots to create more complex charts and package them into a Chart class that can be reused in your own strategies and backtests.