Blame John Walkenbach if you don’t like this post. There, I said it.
He started the 14 basic skills for men. And then added 14 basic skills for women. Not stopping there, he went ahead and added 14 basic skills for dogs.
Debra added her 14 cents by writing 14 basic skills for excel users.
I got jittery and quickly searched 14 basic skills for people writing 14 basic skills posts on google. Alas! nothing found. But being the inveterate non-give-upper I went ahead and prepared my list.
<drum roll> here is the,
14 basic skills for people making charts (or graphs or plots or ok, you get the point)
- Know when to not make a chart
- Know the basic chart types – bar charts, line charts, scatter charts, bubble charts, pie charts, dot plots and more
- Know how to massage, pre-process, arm twist your data so that charts can be constructed
- Know formulas like vlookup, match, index, indirect so that you can get what you want
- Know what chart junk means and how to get rid of it
- Know the types of charts to avoid and the perils of using them
- Know how many and what colors to use. Understand concepts like contrast and repetition
- Know that keeping it simple is far more effective than keeping it complex
- Know that charts are stories and give priority accordingly.
- Know how to use chart templates
- Know charting concepts like dynamic charts, sparklines, tag clouds, combining 2 chart types etc.
- Know the limitations of tools and use the right ones for the occasion (excel for simple and easy graphs, manyeyes for complex visualizations, R for fun and elaborate stuff, Google spreadsheets for maps and organization charts etc.)
- Know that there are beautiful examples around us and learn from them [what your zoo can teach you about visualization]
- Know where to look when you are stuck – PeltierTech, PHD Charting Pages, Google.
Now, it is time for the big screaming question:
How many of these skills you have? Where do you think you should improve?
Please share with us below.
11 Responses to “14 Basic Skills for Chart Makers (Big question: How many do you have?)”
And according to Stephen Few (in his latest book), you need to be interested, curious, self-motivated, open-minded, flexible, imaginative, skeptical, aware of what's worthwhile, methodical, capable of spotting patterns, analytical, synthetical, familiar with the data, skilled in the practices of data analysis. That's 14 too!
And I wrote a while ago 14 Misconceptions About Charts and Graphs
Strange...
Hi Chandoo,
Have used ManyEyes and found it really useful. However, for some confidential data, such web-based third party apps are ruled out.
Are you aware of any desktop based application (free / open source would be ideal), that can perform similar visualization or work on unstructured textual data?
Googling gave me TokenX but wanted to see if you / others reading your blog had a better alternative.
Cheers,
Sumit
@Jorge: And we posted this on 14th july. Too many coincident now... Btw, thanks for sharing the ideas from Few's book. I havent read it (yet). I found your post pretty interesting and sound.
@Sumit: I have used java programs to analyze unstructured text. I used the programs to first structure / clean-up the data and fetch information (statistics, answers to my hypotheses, trends, frequencies etc.) Then I used Excel to analyze / plot the information.
But that was really long back. It has been a while since I did any sort of text analysis work.
Question for other commenters: What tools (if any) you use for text analysis?
...and I would add knowing the work arounds to try when VLookup doesn't work correctly, you know its a data format problem and your not sure why it's not working properly. This @#!?!!@#!! happens to me about once every 6 months....wasting time.
Does anyone have a nice systematic practice you follow to deal with Vlookup mismatch issues that aren't apparent and don't respond to the usual fixes?
Hi Chandoo. Great post, and there's nothing on that list that isn't essential.
I'm going to sneak this one in at number zero: "Design is... crafting communications to answer audience needs in the most effective way." From a great post at http://www.juiceanalytics.com/writing/think-designer/
I have to point out that many things on your list aren't basic skills. Many are advanced skills. Even the more basic ones can be considered advanced in combination. Anyone with all the skills you've mentioned above is a guru in my book, or well on the way to becoming one...because you have to really know your tool (Excel) AND your message AND your audience AND your analytical craft to get it right
Even just your point number 3 (Know how to massage, pre-process, arm twist your data so that charts can be constructed) might require not just a good understanding of Excel, but also of the underlying databases you are getting the data from...and most probably a little understanding of the query language you must use to get at and suitably structure the data you want in the first place. Because to paraphrase Tufte, if your charts aren't interesting, then you might well need to get at some different data.
(Aside: I've recently discovered the power of SQL, which after learning from scratch over the last 5 months allows me to do things that I could never do in Excel given the amount of data involved - or things I wouldn't want to do in Excel even if I could. In fact, learning SQL is the best thing I've ever done to improve my Excel outputs.)
But even limiting your list to just Excel, then to do number Three then you're probably going to need to know not just about...say...pivottables (which can't be that basic, given way more than 90% of Excel users haven't even heard of them) but much more advanced things like how to say cannibilise a GETPIVOTDATA function so you can pull out the specific bits of data you need based on 2 or 3 reference variables listed outside of the pivottable itself. Not to mention, you've got to understand constraints with pivottables themselves (such as grouping a pivottable affects all other pivottables that share the same pivotcache) , and how to get around them.
Even just this small part of your point number three on it's own is advanced.
THe good news is that (as Stephen Few points out in his book, and as Jorge comments above) IF you are interested, curious, self-motivated, open-minded, flexible, imaginative, skeptical, aware of what’s worthwhile, methodical, capable of spotting patterns, analytical, synthetical, familiar with the data, skilled in the practices of data analysis...then you'll be able to pretty much well cover off the 14 PHD points above after a year or so of exposure to Excel coupled with exposure to some pressing data and visualisation problems you need to solve using it (nothing improves your understanding and skills better than a challenge to sharpen your claws on).
No, wait...that's the BAD news. The GOOD news is that we can see forther by standing on the shoulders of giants.
[...] material for reading: 14 basic skills for chart makers, Make dashboards using [...]
Hey Chandoo,
Am here again!
So, when we talk about charting some data there is 1 thins which one should do before actually charting the data.
before any starts building charts out of data, one should understand the chart and the statistic that is being represented on the chart. For example, if one is plotting a market share data, then he would require Pie chart, whereas if the same data is being shown on a time-series, then a 100% barchart would be required.
The other thing is knowing the statistics. Please do not get confused here. I do not want to say that one must know statistics. What I am trying to say is that one should know what type of data one is plotting, is it number of... or is percentage of...
This is one of those skills which I have seen that many people really do not know, or do not pay attention to in case they it. Another example is the scale of chart. Visual statistics can be as mis-guiding as you can make.
Let me know guys what you think about this.
[...] 14 Basic Skills necessary for Excel Chart Makers [...]
[...] Charting Skills for Dashboard Makers [...]
Dan - I find the use of "iserror" and similar formala invaluable when not only graphing but in all work. I choose to display " " instead of error values - some choose zeros etc.
An clean/ sort data as a first step in every process or project. This may help in reducing mismatches.
Hope I'm not telling you how to suck eggs... but these little steps are so valuable in all that I do.
[…] 14 skills you need as a chart maker […]