search

TomorrowsCities / tomorrowscities

MIT License
3 stars 1 forks source link

Zero values for impact metrics #36

Open karimaljawhari opened 3 months ago

karimaljawhari commented 3 months ago

Tried using the scenario of rain-induced landslides for Nablus. Impact metrics show zero values. Could be an issue with loading the full content of exposure database. Input data and analysis log are attached.

test_nablus_landslide.zip

hkayabilisim commented 3 months ago

Hi Karim,

Short version: If you remove the last two fields in "expstr" of landslide fragility, it will work. For instance: Original version:

image

New version:

image

Long version Source of the problem: The problem is related to the incompatibility between "expstr" is landslide fragility and "expstr" in building layer. The "expstr" in landslide fragility is composed of material type and code level such as "Adb+LC" or in your case "CR+CIP/LDUAL+DUL+HC". However, in you landslide fragility excel, you specify height and occupancy type.

Why: You may wonder why we don't create a 1-to-1 mapping between expstr columns of building and landslide fragility. The reason is that, it will require to include all unique exposure strings in the fragility. This means lots of repetitions. In Rapti study, the team suggested that the susceptibility only depend on material and code level. So we decided to use "material type + code_level" in expstr of landslide.
There is some explanation here but they are not clear enough.

Quick solution The fast solution is to get rid o the last two fields (height and occupancy) in landslide fragility. In this way, without and change in webapp, you can run landslide computation.

Another solution We can change the engine so that the expstr of building and landslide fragility must match. In this case Rapti landslide fragilty data should be regenerated.

My suggestion You go for the quick solution. In the meantime let keep this issue open, and discuss this further in future.

hkayabilisim commented 3 months ago

And make sure the combination of expstr and susceptibility in landslide fragility is unique . Added @jamaldabbeek

hkayabilisim commented 3 months ago

Added a new pre-compute check related to this issue: https://github.com/TomorrowsCities/tomorrowscities/pull/38

hkayabilisim commented 3 months ago

Can you try with this fragility: vulnerability_landslides_update.xlsx which is created by removing redundant rows.

jamaldabbeek commented 3 months ago

thanks it is working