matplotlib
Feature Request: Plot percent price change on the y-axis
Is your feature request related to a problem? Please describe.
When I want to compare two stocks, I'm interested in seeing the % changes in those two stocks over a period of time. Currently when I plot two stocks in mplfinance (using a secondary y-axis), I can't see the day-to-day percent changes, I only see the day-to-day price changes, which can be on different axes and thus inaccurate to compare directly.
For example, if I want to compare SPY and GOOG over a time period, I want to see the percent changes between the two (for easy comparison) rather than their stock prices which are almost an order of magnitude apart.
Describe the solution you'd like
A plot option that lets you plot the % change and lets you specify the start value that the percent changes should be compared against.
Hi @fintron
I'm having some difficulties understanding the feature you are requesting. Are you looking to plot two plots with the percentage difference from a specified value for each stock?
-- Caleb
In one chart, I want to plot two series. The y axis is the change each had from the start of day.
Does that make sense?
So just to clarify, you want the percent change from the Open to the Close each day (as opposed to percent change from previous day's close).
So for example, if this is your data:
| Date | Open | High | Low | Close |
|---|---|---|---|---|
| 2016-03-04 | 102.370003 | 103.750000 | 101.370003 | 103.010002 |
| 2016-03-07 | 102.389999 | 102.830002 | 100.959999 | 101.870003 |
| 2016-03-08 | 100.779999 | 101.760002 | 100.400002 | 101.029999 |
| 2016-03-04 | 101.309998 | 101.580002 | 100.269997 | 101.120003 |
The percent change would be:
| Date | Open | High | Low | Close | PercentChg |
|---|---|---|---|---|---|
| 2016-03-04 | 102.370003 | 103.750000 | 101.370003 | 103.010002 | 0.625182 |
| 2016-03-07 | 102.389999 | 102.830002 | 100.959999 | 101.870003 | -0.507858 |
| 2016-03-08 | 100.779999 | 101.760002 | 100.400002 | 101.029999 | 0.248065 |
| 2016-03-04 | 101.309998 | 101.580002 | 100.269997 | 101.120003 | -0.187538 |
Is that correct?
taking the first row, for example, that was calculated as:
100.0 * ( (Close/Open ) - 1.0 )
or
100.0 * ( (103.01/102.37 ) - 1.0 )
So the plot should look (for example) something like this:
Is that what you want?
No - actually I meant that I would pick a value which is the baseline - for example "102.370003". Then, I will just look at the percent change from that value for the remaining days. Does that make sense? The user could for example choose 2016-03-01's open value and then basically see the percent change from that point.
This is useful to say plot the % change of the SPY over time vs a stock you are interested in.
Do you have an example of such a chart from another location? If not, no worries, I will try to work up a prototype from your description and post it here to see if that's what you are looking for. But if you can point to an example that would be helpful.
Here's an example plot from Yahoo Finance showing what I had in mind - GOOG vs SPY:
https://finance.yahoo.com/chart/GOOG#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%3D
No - actually I meant that I would pick a value which is the baseline - for example "102.370003". Then, I will just look at the percent change from that value for the remaining days. Does that make sense? The user could for example choose 2016-03-01's open value and then basically see the percent change from that point.
This is useful to say plot the % change of the SPY over time vs a stock you are interested in.
@fintron It appears to me, from the example tahoo finance plot you posted above, that rather than specify a specific baseline, the plot always chooses the first date (first data point) as the baseline. That makes sense to me. Any concerns or comments?