makella
commented
5 years ago hi @elenatorro and @LadyAga! figured I'd bring the conversation about the Quickstart Guide over here!
from the doc:
question here: for the quickstart, should we go straight to the helper instead of adding all things, removing them, and then using a helper? alternatively, we could put the helper step after the "analysis with PySAL" and visualize that layer with a helper
Yesterday we discussed with +elena@cartodb.com that it would be better to avoid using Helpers in this Quickstart guide (to keep it as simple as possible). What are your thoughts on this, +makella@cartodb.com?
I can see this both ways and wanted to discuss that a little more!
in my mind, helpers are meant to be a quick start for visualizations (that include widgets, popups, etc.). so I worry that diving into some of the details like this in the Quickstart will make CF feel intimidating.
On the other hand, current helpers are most useful if you want to create a data-driven visualization. The other way of seeing it is that we are missing just a default_layer helper where you aren't styling on an attribute but instead have a dataset that you want to explore and be able to add a widget , a legend, popups, etc. We have all the pieces to put that kind of default layer helper together and i'm wondering if we should think about adding something like that? I see where the complications will arise in enabling the option to choose the value for each (for example, during the exploratory phase you may want to add a category widget on one field and a histogram widget on another similarly the legend, etc.)
For me, going straight to the legends, widgets, and popups at the layer level and then styling the final map with VL syntax, in the Quickstart, wouldn't demonstrate how "easy" we've made things! And this is a place where I'm guessing a lot of people will start.
Since the default_layer idea would be additional work, here is an idea for an alternate outline:
- Explain what we are going to do
- Explain what the dataset contains and allow to download it
- Log in to CARTO
- Geocode the file
- Show it in a bare map (no legend, no popups, no widgets)
- question: what will the point dataset be? if they are just meant to aid in the analysis, I don't think we need to add too much information about them on the map.
- Instead of adding anything to the map, we could show how to look at the attribute table, and explore the data with the SQL API client things like (
distinct, count, bounds, create_table, insert_table, rename_table)
- Add a second layer with no car households data
- Visualize with
color_bins or _continuous_layer - Analysis with PySAL
- Upload data to CARTO
- Make adjustments to final map inside of helper methods (add widget, adjust legend title, description, adjust style, etc.)
- Publish map
Again, this is just another take on the same thing and I think this workflow would be less complex, but look forward to discussing our options!
newLadyAga
Related with: Documentation Refactor EPIC
Context
The Quickstart guide should be a simple example that contains the main concepts we want to show to the first time users at a glance.
Current Status
We've decided to split the Quickstart Guide into 2 parts:
In the first part, we'll read data from different sources and we'll create a visualization using a helper method
In the second part, we'll use the data observatory and the data enrichment
This is currently blocked by https://github.com/CartoDB/cartoframes/pull/1083, cause the polygons enrichment method is failing and we need this PR to be merged to continue with this task.