Just back (to the basics)

Average and colors

This pause, perhaps longer than planned, offered me a chance for chatting with a customer about the importance of the correct data management.

The nub of the matter was about “undervaluation” of the use of the basic statistical functions.

Reviewing the toolbox

These are the “tools” that cannot miss in any practitioner. A useful notice: using discrete values (e.g. considering that time is fractioned into single units) makes much easier to produce graphics supported by formulae that can be sold with no or little introduction.

1. mean
2. mode
3. median
4. Average absolute deviation
5. Standard deviation

Travel and weather

Most of the statistical manuals use the temperatures for showing the possible pitfalls of an oversimplified usage of mean. Although this value could easily sidetrack any evaluation, it remains a “necessary but not sufficient” tool for understanding the trend.

In the following table, there are “described” the climate differences between two cities lying on the same latitude.

For a short business trip, a quick glance to the weather forecasts covers most the logistic needs. A leisure journey requires much more details. However, planning a big building project (near a river) needs much more information. Consequently, the decision about allocating the necessary resources for processing the information is dominated by the influence of each outcome and then output.

Who withholds what

Just for compiling this simple table, not all the data were immediately available. The ever present Wikipedia (Naples) does not show the average for rainy days. Nevertheless from a more serious analysis, this information would be considered acceptable without including other data.

The quest for the correct information shall be planned having, as a compass, the scope of our search.

The risks related to the incorrect search are numerous and bigger than normally expected:

1. Wrong focus(es)
2. Too many details forming a halo that hides the reality
3. Excessive costs and delays that are difficult to be explained without meaningful results
4. Rousing warning (either about turf invasions or incompetency)
5. Missing the necessary links with the other pieces of the puzzle.

Simple tools (for the beginning)

Information is meaningful only if it is properly framed into the correct scenario. Once this operation has been done (it is always a team-work), the second step is about the understanding the desired impact (opportunities or risks).

Like any decision, this too can be subdued by biases. At the beginning of the process, the usage of simple tools, with the provision of the preliminary phase, offers good opportunities for a quick and easy to share analysis.

The importance of meta-data

As soon as the guidelines for the search are set (and accepted by stakeholders), the next step is to establish (with the help of SMEs) the necessary meta-data. These have to reflect the organization’ structure for many reasons (including the biases):

1. Who produced and checked it
2. When / Where
3. Which is the used method for collecting data
4. What it is going to be used (in relation with the size of volume and the complexity of the operations required).

Conclusion

Statistics have a dreadful reputation. Using them wrongly is a painful sort of professional suicide. However, crunching (correct figures) is an essential part of any manager. In a project, this shall be done carefully to avoid the creation (at least in the beginning) of unpleasant rumors. From a risk management viewpoint, the right (and updated) figures make the difference between success and failure.

More readings

http://online.wsj.com/article/SB125200842406984303.html