Visualise large time-series data

Hi, I have a big file of time-series data. The problem is when I use Plotly graph object to visualize, it is very slow and worst when I try to zoom in (using rangeslider) , which is important.

So, every time, I have to take a window of data from pandas dataframe and visualize it. This is not very convenient.

So, is there any way, that, for the first time it will show the first window of data and when I scroll, only then does the rendering of the second window happen (from memory or even from hard disk) and so on…

I know, with Dash it is possible but is it possible in Graph Object?

Hi Mainul,

Indeed, it is not that straightforward to achieve this using solely Plotly Graph objects, as these store all the data in the javascript-figure.

I suggest you take a look at GitHub - predict-idlab/plotly-resampler: Visualize large time-series data in plotly, this repo creates a dash-app (which can be still be shown inline in jupyter/colab/vscode) and is able to visualize huge amounts of sequence data by using computational aggregation.

Hope this helps you!

Best regards,
Jonas

2 Likes