- Real-time visual analytics allows exploring data visually and interactively to gain insights. This is more powerful than static reports as it uses human visual perception to see patterns and relationships.
- In an example, a sales manager is able to visually analyze sales, expense and profit data for a beverage company across states to see how different products contribute to profits in different areas.
- The analysis is done live and interactively to uncover insights like certain products being unprofitable in some states and relationships between sales and profits for different vendors. This interactive visual analysis provides a better way to work with data than delayed traditional reporting.
2. Visual data analysis lets us use our powerful visual
perceptual abilities (we all have this) to quickly and
clearly see what our data is telling us
Doing it in real time lets us think our way through the
questions and the insights that the data reveals
PS: This is the reading version of this presentation. Never present and speak to
something with this much text!
3. Let’s illustrate this with an example. Suppose you are the sales manager for a hot
beverage company. You have data on your sales, expenses and profits across the United
States. You know there are some disappointing overall results in several areas, but also
that it is not clear what is driving these. But the Board wants to know. Having fun yet?
4. Here’s part of a set of data, representing sales, expenses and profits for your
beverage company.
Would you rather be able to see the trends and relationships in the data
right now? Or would you prefer to ship it off to the data analyst and get a
technical report later?
* This hypothetical data is supplied by Tableau™ software for product demonstration and education
6. What if you could explore your data live with your team,
and then take your management, again live, through the
findings that you have uncovered? Instead of trying to
speak to a report that you didn’t even write.
7. Here’s a first look at the data, analysed live. This took virtually no time
whatsoever, using software capable of instant data visualization. So let’s dig a
little deeper – we want to see where these sales are coming from and why some
are so high, or low……
(the data graphics in this presentation were produced with Tableau Desktop™)
8. Again, dragging and dropping in a new variable, product type, we can see that
the picture is not so simple … the primary sources of sales vary across the
states… it’s a bit messy, so let’s sort on one of the categories….
9. There we go, sorted instantly. Now we see more clearly some of the key differences
in the sales of the product types across the states by simply looking at the shapes of
the distributions relative to the sorted one and each other.
So, how’s this related to profitability? Do profits mirror sales?
10. Not exactly a one
to one relationship
between profits
and sales by state
– at least not in
some cases.
So there is lots to
look at. Suppose
we look a little
closer at what lies
behind these
aggregates?
11. We see lots of things. Highly profitable California is so due to strength in espresso
and teas, as is Illinois. We’d have a more profitable operation in Missouri if we didn’t
try to sell herbal tea there, nor espresso in New Hampshire. Little profit is made in
coffee in Iowa, nor Nevada, which are overall high profit states. And so on…. visual
analysis works quickly and powerfully.
We could also visualize this as a table…..
12. Here, use of colour coded to profits makes the points of
interest stand out.
13. Suppose we are the coffee team. We are surprised by results such as
California, where coffee has not been the profit driver.
Here we can produce, visually, the cross tabulation of profits, sales and expenses by
state, for the three types of coffee we sell.
For instance, we see clearly that in California, we are only doing well in Columbian
coffee, thus reducing “coffee” overall as the profit driver in that state. Expenses are
too high relative to sales for Amaretto and Decaf Irish Cream. No need for
econometrics here!
14. Remember, all of this analytical power and clarity has come from data visualizations
that can be done on the spot, at the monthly sales meeting, from this unfriendly –
looking mess of data (this is just the profit data) below. No need to send away to
the data crunching department to find out what’s going on.
15. Putting key views into a dashboard, where everything can be seen at once allows even
more analytical power and clarity. This, again, is done on the spot during visual data
analysis.
16. Perhaps we’d like to see a more direct analysis of the relationship between
sales and profit. A quick drag and drop and we have a scatter plot with
trend line. It’s a pretty strong relationship, but with a significant set of
unprofitable outliers. The points are individual vendors.
17. To look at this in a bit more detail, we can use a “trellis” or “small
multiples” display. We see here, as before, that there is a problem with
two of the products in the West. Since we already know that California
is where this is happening most, let’s go there…..
18. With another quick adjustment of the visual, we can see that the relationship of sales
to profits, while poor in both cases, is different for the Amaretto and Decaf Irish
Cream products – in the latter case, the more is being sold by a vendor the more their
profitability falls. These would offer different implications to the product manager as
to how to improve these results.
19. In this demonstration, visual data analysis let us use our
own powerful visual perceptual abilities see what the
data was saying. And this could have been done live, at
the sales meeting. No reams of statistics and charts, no
delay.
That’s a better way to work.