Now, even though I have looked at the long-range forecasts from half-a-dozen different metrological organisations, I wouldn't be surprised if this prediction is wrong. That is because the complexity of the weather means that even relatively short-term forecasts are far from reliable, a fact which explains meteorologists' fondness for 'nowcasting'.
Nowcasting uses good old-fashioned human beings to assess the latest radar, satellite, and observational data to predict the weather for the next six to 12 hours. It is only a small step forward from feeling seaweed and chanting "Red sky at night", but, with reported 80%-90% accuracy, it is state-of-the-art.
So meteorologists forecast tomorrow with a high degree of accuracy - but economists will tell you they have the even tougher job of forecasting for today with similar precision. The UK's second quarter GDP growth forecast of 0.2% was announced three weeks after the quarter had ended, whereas GDP figures for Q2 won't be available until October.
Which explains why economists are also interested in nowcasting - and why they have been looking particularly closely at influenza.
Seasonal 'flu epidemics are unpleasant, disruptive and result in up to half a million deaths each year. Early detection and rapid response to outbreaks can significantly reduce their impact. But herein lies a problem, because the data recording the number of patients presenting influenza-like illnesses (ILI) is typically gathered on a weekly basis, with a reporting lag of one to two weeks - and in this time-lag a 'flu outbreak can run riot.
However, in 2009, statisticians at Google published a letter in Nature offering a simple nowcasting solution. "Because the relative frequency of certain influenza- related Google searches is highly correlated with the percentage of physician visits in which a patient presents with ILI, we can accurately estimate the level of weekly influenza activity in each region of the US, with a reporting lag of one day."
One day! That is a game-changing difference in itself, but what is even more important here is that Google searches are a "precursor signal": patients are looking for information on 'flu symptoms, a precursor to actually coming down with 'flu and visiting the doctor. In other words, the Google team hadn't just uncovered a powerful real-time nowcasting tool - they had found a nowcasting tool that was running ahead of real-time.
Economists pounced on this discovery and are using the same technique to predict the movement of key indicators, such as retail, car and house sales. All of this came from data which was already being captured and which is freely available.
Which brings me to HR - because in my column last month I highlighted the CIPD's 'Next Generation HR' research project and the pivotal role it places on an insight-driven approach to people practices. I observed that most HR departments are awash with data: it was identifying insights buried within the data that was the real challenge.
This is why I am excited about the nowcasting concept, because it gives us a valuable lens through which we can look at data in a new and interesting way.
What are the things we want to predict about our workforce (engagement, turnover, uptake of benefits etc)? And are there any precursor signals in our existing data that will enable us to be ahead of the game?
Imagine the impact HR could have on an organisation with insights like these.
As for this month's weather, we just have to take what we get and remember the wise words of Billy Connolly: "There is no such thing as bad weather, just the wrong clothing, so get yourself a smart new raincoat and live a little."
David Fairhurst is chief people officer, Europe at McDonald's
