Challenges with pydeck reading csv

Using Streamlit
1

Hi Streamlit Community

Still getting to grips with Stramlit, but getting better. However, pydeck_chart has been giving me a few headaches. Namely, whenever I try to read a CSV I get an Unexpected token N in JSON at position 53116 error. The script will work fine on a CSV which only contains lon and lat columns, but when I add further columns to the CSV, suddenly the script fails citing the above error.

I have looked through everything I can find but still feel at a loss. Is there a specific way one should be formatting the csv?

I have attached a link to the repo that contains:

  1. the script (adapted slightly from the columnlayer tutorial)
  2. the basic version of the csv with only lon/lat data
  3. the expanded csv with the same lon/lat data, and a few more columns of number. No blank or null entries.

Any help much appreciated, even if it is just pointing to a tutorial on this topic. Apologies if this is very basic or not streamlit specific.

2

Check your data and the dtypes after your imported it with:

df.info()

I guess you have some null values in your data and you have to convert the object types explicitly to float, string or whatever the proper datatype is.

3

Thanks again Franky, this helped put me on the path to fixing the issue. The main errors were:

a. a calculation had been done on empty cells, giving a divide by zero error code
b. there were a handful of null values in the table.

Once these were resolved the layer was rendered.

4

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.