Non-Guided Project: Making An R Shiny App to Track Moths

Hi folks!

I took a bit of a detour from Dataquest to complete the Our Coding Club Data Science for Ecologists course and got really excited about the Shiny framework for R, which allows you to make interactive web apps.

I created this interactive web app that uses data from the app iNaturalist to track observations of Luna moths across the US (not very mobile compatible, sadly - I’m not a CSS pro yet). Let me know what you think! https://phoebethatch.shinyapps.io/lunadash/

Have any of you guys had experience with Shiny?

6 Likes

Hi @phoebepreble

Shiny is pretty cool. But I guess it doesn’t really cater to people, who are just learning data science. I mean you need to deploy it on a server, right? A lot of additional prerequisites here. However, I like what you came up with. Maybe one small caveat. I guess selecting the years leads to cumulative sums being visualised. Is this correct? From the interface this is not really obvious. Maybe adding a note somewhere would already help.

Btw: Are you involved with this peer-learning community? Looks pretty great.

Best
htw

Hi htw - re:Shiny, I think it’s actually pretty accessible for beginning coders! If your computer can run RStudio you can probably test apps using a local server (the package comes with this capability, i believe).

hosting them online is another matter but shinyapps.io lets you host 5 for free which is very reasonable for students.

yes, the year UI doesn’t really make sense, especially with the time limits of the data - I’m thinking I might just make a long timeline containing all 4 years instead, do you think that would be a more intuitive system?

@phoebepreble

I guess you are right. It does make a difference if you look at shiny as a learning opportunity or if you need to care about the server deployment. Thanks also for pointing out that you can host 5 apps for free at shinyapps.io - I wasn’t aware of that.

Re: Ui. For me, your suggesting with the long timeline does make sense. Or at least that what would feel natural to me.

BTW: Are you only coding in R? Or also in Python? In Python there are actually nice visualisation options for handling interactivity now as well.

Best
htw

2 Likes