CenterForTheBuiltEnvironment / clima

The CBE Clima Tool is a web-based application built to support the need of architects and engineers interested in climate-adapted design. It allows users to analyze the climate data of more than 27,500 locations worldwide using the data contained in EPW files.
https://clima.cbe.berkeley.edu
MIT License
61 stars 21 forks source link

Datetime filtering #119

Open praftery opened 3 years ago

praftery commented 3 years ago
praftery commented 3 years ago

One additional point regarding the caption text that explains that a plot has been filtered. It would be better if this was a plot annotation, or something similar so that when a user downloads the svg image, that filtering information is still on there.

giobetti commented 3 years ago

Several points here, so I am opening separate issues for those.

One is to make sure to review all graph titles and make sure that the title is part of the downloadable graphic to avoid confusion #122

As it currently stands, if someone is interested in a specific period of time each day (e.g. 9am-5pm), they have to define and apply that filter on every plot they want it on. It would be easier/better to do this once, when needed. Consider adding an option to apply/change the filter across all relevant tabs. I believe this would be quite cumbersome to implement. Also different graphs might need different time filters (i.e. UTCI and natural ventilation) Also, this is sort of implemented already in the "Data Explorer" tab in the "Customizable heatmap" where you can set the time and variable filter and those will be kept even if you change the variable to be displayed.

Additionally, in plots where more visualization dimensions can clearly be added without overloading the plot, it may be better to combine filtered periods with the unfiltered annual data, rather than just dropping the unfiltered data The disadvantage here would be that it woul be difficult to answer the question "what percentage of the selected time x happens?" while now it is trivial.

In plots that inherently show one of the filter values as a dimension (e.g. heat maps), maybe it's better to keep the data and just add horizontal/vertical lines (on apply a partially transparent mask to grey it out)? It would save on processing/data/load times etc and seems a bit more intuitive that completely whiting out the plot. We experimented with this and where not really satisfied with the grahical output (often very confusing to read) It might be a matter of tast but I find it way clearer as it is.

Maybe allow filtering by datetime features for the descriptive statistics table in the data explorer? Could be a good enhancement opened as #123

As with the other plots, the psychrometric chart needs to have an annotation stating when it has been time filtered (to prevent user error on download) part of #122