Housing Minister Grant Shapps has an exceptional quality. He sees sunshine when all about him see black clouds. That’s an uplifting and admirable quality, up to a point.

That point appears to have been reached.

His sundrenched interpretations of what many see as gloomy official data seem to have created something of a storm. His distinctive use of statistics has even spawned twitter hashtags such as #shappstistics.

Certainly it’s very worrying when those determining policy put an overoptimistic gloss on important data that blinds the public to nasty realities, or selectively choose which data to release. I have discussed this before.

But in fairness to Mr Shapps, until recently his shadow, Jack Dromey, appeared equally willing to trade in what many expert observers might have deemed selective interpretation of data.

To some extent this is expected of politicians. However, in the middle of a housing crisis it might be deemed unseemly if not exceptionally inappropriate for a Government minister.

The game, however, has moved on. Mr Dromey reported Mr Shapps to the statistics regulator and brought the issue into the wider media spotlight.

While others will discuss the politics of this move, it is worth looking at the issue of politics and statistics.

For me the central problem is the reduction of statistics used to describe complex things to the direction and distance between two data points.

This has degraded statistics and data analysis within politics and the media to a simple notion of better or worse – some might suggest the stock in trade of both professions.

Better and worse are by definition relative and if we reduce statistics to the difference between two points it is almost inevitable (given a suitable time series) that we will be able to find two convenient points that will support the argument we are making however bonkers that argument might be.

The point about statistics, why they developed, is that they help us better understand the world. And they can be pretty sophisticated and very useful.

So, for example, for most people most of the time the average is the total of a group of things divided by the number of things in the group. But statistics has an array of different ways to express an average or, put another way, to describe a central tendency.

The general public, for instance, may not be regular users of the geometric mean, but they feel its impact through its use in calculating the CPI and not the RPI measure of inflation.

That level of sophistication may seem excessive. But statistics isn’t about absolutes, black and white, up and down. It is a developing language to help explain things.

It is about uncertainty, probability, tendencies, trends, turning points, distributions, variations, proportions, correlations, confidence, significance and a host more.

Furthermore there are both numerical and visual ways to express the findings to help improve our understanding.

To reduce it solely to which data point is higher or low than another data point is, frankly, barbaric.

For me the problem does not lie with statistics. In many ways it doesn’t lie with the politicians. Ultimately the problem lies with a public and media (on the public’s behalf) that allow politicians to distil data analysis to the distance and direction between two questionably selected data points.

The media should be far less willing to allow politicians the space to quote statistics unchallenged and that challenge should not be met simply through pitting against each other two politicians spouting different differences between two different data points.

It does the public no favours, politics no favours and does no favours to the policies that eventually shape our lives.

Data and statistics are becoming an increasingly important part of the news media mix. The media must respond and to find more appropriate ways to convey and challenge statistical information.