tessellation not recognized as a gpd.GeoDataFrame

[MehdiAbarro]

Hello,

I have an issue creating a FlowDataFrame: my tessellation is not recognized as a gpd.GeoDataFrame.
Here is how I define my tessellation and my flow data:

Then OD = skmob.FlowDataFrame(data=OD_flows, tessellation=tessellation, tile_id='tile_ID') raises : "TypeError: tessellation must be a GeoDataFrame with tile_id and geometry."

[me9hanics]

Because the thrown error is TypeError, I believe the issue may be that one of your columns has a different type than required. At the end of your code, check the data types with tessellation.dtypes of both columns, first should be int64, second should be geometry. It may even happen that it has issues with OD_flows and the stored datatypes (I had these issues just yesterday for a created FDF), in that case, do this:
fdf['origin'] = fdf['origin'].astype('int64')
fdf['destination'] = fdf['destination'].astype('int64')
fdf.tessellation[constants.TILE_ID] = fdf.tessellation[constants.TILE_ID].astype('int64')
(constants come from from skmob.utils import utils, constants).

Maybe not having a coordinate reference system for the GeoDataFrame causes the issue? Might have problems with "interpreting" the geometry. Try tessellation.crs="EPSG:4326" (or even better would be taking the CRS of ch_geo) after turning it into a GDF in line 2.

One last thought is that maybe you might also need a population column, from the example here. If that is the case, the developers should make it clear when throwing error.

[MehdiAbarro]

Thank you for your answer!

Sadly, even after checking everything you mentioned I am not sure why the error raises.

I still have the same error raised when doing OD = skmob.FlowDataFrame(data=OD_flows, origin='origin', destination='destination', flow='flow', tessellation=tessellation, tile_id='tile_id').

What is even weirder is that when I try to run the provided example the same error raises.

[MehdiAbarro]

Still have the same issue, not sure what I should do...
When checking the logs I noticed that this is the snippet where the error is raised:
`if not isinstance(data, pd.core.internals.BlockManager):

        self[constants.ORIGIN] = self[constants.ORIGIN].astype('str')
        self[constants.DESTINATION] = self[constants.DESTINATION].astype('str')

        if tessellation is None:
            raise TypeError("tessellation must be a GeoDataFrame with tile_id and geometry.")`

So my geodataframe tessellation is considered to be None even though it clearly is not. The default input for the tessellation parameter is None so I guess the FlowDataFrame class ignores my geodataframe ?