I have some issues plotting "large" datasets of timeseries data in python, where the time jumps across a few decades in erroneous samples. We aim to visualise only the timestamp (unixtime + custom microseconds) vs index. In this example there are roughly 40k samples.
Basically, I am assuming it is some issue with the rendering of the plot by matplotlib, because when I move the axes, both the scatter points and also the lineplot seem to glitch all over the place. A further bit of evidence for this is that the line in the lineplot is not actually going through the markers, when I zoom in or pan the plot.
The timestamps are continuous and increase by 40ms between steps.
Overview of errors (timestamp is zero -> default date 1.1.1970)
Example of how the plot should look like
Timestamp raw data (ignore ms fraction 2)
Code used to plot (using google colab, re-created in Visual Studio Code)
if single_file_or_multiple == "multiple":
fig = px.line(df_trace, x=df_trace.index, y="time", markers=True,
color="rec_id")
fig.show()