Charting for fun and profit

by Admin
Updated: July 3, 2018

Charting for fun and profit begins with using visual processing to reveal patterns and trends

Charts can be useful because they allow you to find correlations using the parts of your brain that processes visual data. This is visceral and fast and evident in language - when we understand something we say, “I see.

If someone throws you a ball, you can catch it because your brain predicts where it’s going using visual processing, i.e. no conscious use of mathematics required.

If someone throws you a ball in the dark it is much more difficult to catch, even if you can resolve vectors & forces in your head in real time (difficult but not impossible).

In an analogous way, GPUs are better suited than CPUs for mining crypto.

What is important

It’s possible to create charts to support any idea. In order to protect against spurious correlations, it is always necessary to think very carefully about the underlying data.

  1. What you measure
  2. How you measure it
  3. How you display it

Example: Suppose you want to investigate seasonal variations in business.

Firstly, you need to decide if you are counting income or profit.

Secondly, you need to decide on basis (cash or accrual) and sampling rate (quarterly, monthly, weekly).

Thirdly, you need to render the data in a meaningful way (in this case chronological order on a bar chart is probably best).

When you have the results, it’s a matter of creating other charts of possible causes and seeing if they match the pattern, e.g. a peak in summer might relate to outside temperature or it could be because of sales efforts in the spring (patterns match but are out of phase).

Quick chart tool

Enter data in the form, descriptor = number one per line for a horizontal bar chart (it doesn’t matter if there are spaces either side of the equal sign). Sample data is provided on page load. You can paste or load* your own:

ff

* Browse & load is not upload - it uses JS FileReader so your data is loaded directly into your browser only and not sent anywhere via the Internet.

This is a quick & simple tool for analyzing data - it took only minutes to write. I’m posting it here, buried in an article because, if I made it a formal tool, I’d end up spending weeks refining it when plenty of better alternatives already exist... but I respond to requests 🤓

Patterns vs trends

Patterns are repeating cycles such as seasonality. Trends are progressive such as year-on-year growth.

Charting raw data should reveal both. Normalized data helps show patterns, e.g. monthly percent of annual total. Aggregated data helps show trends, e.g. total annual sales plotted against year.

Monetizing charts

  1. Charting facilitates deeper understanding from which improvements can be made.
  2. Charts also provides visual feedback and motivation. I recommend using bar charts to track your net worth.
  3. Being visual, charts are superior vehicles for communication. They are effective for sales & promotional presentations, motivating staff and attracting investors.

More info

Simple explanation & illustration of phase shift at chegg.com.

“Types of Charts: Choose the Best Chart to Convey Your Message” at mymarketresearchmethods.com.

Charts are most easily created from spreadsheets, e.g. Excel (support.office.com) and LibreOffice Calc (libreoffice.org).

Charts can also be created as Scalable Vector Graphics (SVG) for the web, e.g. css-tricks.com.

Internal links

Correlation and causation Pareto distribution Break-even Candlestick charts Go to Articles
Agree? Disagree? Questions? Answers?
Please post a comment...
Log in