4 ms·
My personal problem with these types of charts that they attempt to present the "raw" data and expect me to sift through it to find the interesting stuff. It is
by jalfresi 11y ago
My personal problem with these types of charts that they attempt to present the "raw" data and expect me to sift through it to find the interesting stuff. It is my understanding that originally bar charts and their like we're used to illustrate visually a specific point from a set of data. This no longer seems to be the case these days. Instead we get an article about "mass migrations of people across the globe" and a visual dump of all the data, like I'm expected to find the interesting story in all this! Why not write about specifics and illustrate that with a traditional chart?
I have similar issues with admin systems or management dashboards (having had to write loads of them over the years). I've found that executive users would much rather see a bar chart even if the bar chart doesn't display data of any significance, than see numbers. They would much rather see a partial non-story be displayed if it meant the visual display and layout of the dashboard was consistent than see two significant numbers. For example, imagine a dashboard that shows performance figures for sales. I've often been asked to show what is usually a static screen layout of bar charts etc that doesn't tell the user anything import rather than show 2 or 3 stand out figures I.e. Sales are down over last month in a big red call-out vs a bar chart that shows all sales figures for all regions compared to last month which requires the user to determine if the overall sales are down on last month.
There seems to be a large disconnect between the purpose of the visual display of quantative data and the action required or message being delivered. I actually put a lot of this down to excel; I've seem countless spreadsheets contain a graph that tells you nothing of significance because the person writing the spreadsheet doesn't understand that there are values of significance in their data. It's like the expect the graph to tell them the story (whilst this is true in some circumstances, I've only really seen this work when the data is obviously significant. Countless times I've see than overall trends can be missed because the use of data is incorrect, and then the wrong type of chart is used to display the data "to look nice in the report")
Anyway that's my main beef with these types of visualisations; they always come across as a kind of "there is far too much data here for us to sift through and find anything of importance so to examine. Let's create a tool that lets th user do the work". Unfair? Probably. Busy? Definitely. Pretty? Absolutely. Useful? Debatable.
- aunty_helen 11y agoI think part of the problem is, when you're working on the data so closely you lose the outsider perspective and things which you think jump out and are relatively easily divulged actually require a deeper understanding only obtained from an in depth analysis. I guess it's a bit of a catch 22.