How to adapt innovation indicator to another city?

[danhntd]

Hi Ronan Doorley et al., @doorleyr @crisjf
Thank you for your very interesting work. I have 2 main issues about the innovation indicator in this project.

I want to compute the innovation indicator for another city (in my case Ho Chi Minh City, Vietnam - HCM). The available data is the geogrid data of the buildings in the area.

1. In the first approach, I tend to leverage all the pre-trained models and source code here to adapt HCM data.
   The only things I change are the input geogrid and the land type definition.
   In this case, I need to align the land type of our data to fit the Corktown land type, then, of course, I have to accept that there will be a semantic gap between the two areas.

The questions in this case are: Why did the Corktown project have specific normalization ranges for its indicators?
And does it make sense to me to also follow these ranges of norm values for my HCM data?

Building indicators: range [50.000,100.000] ([line 152-153])
Innovation Indicators: (line 26-28)

• Skills: range [-16,-5]
• Knowledge: range [-11,-7]
• RnD: range [4,5]

2. In the second approach, I need to train another new model for my HCM data.
   So, in this case, which input data do I need to collect and how can I form those data format to fit into the training process?

Thank you very much in advance.
Danh T. Nguyen