If you’re reading this, you probably know that the official unemployment numbers for August will be announced on Friday. For the past few months I’ve been using Gallup Poll data to input into a model, and make predictions prior to the official announcements; but not this month.

It turns out that I’m getting pretty close to putting something out for a basic circuits tutorial, so I’m going to halt my unemployment model activity, as I would like to get this tutorial going sooner than later. If folks are really interested in continuing to see the unemployment model published, you can always let me know through the “contact us” section, and if there’s some demand I’ll try to keep it going.

Brad

Here’s what I decided.  Once I publish my forecast (2-3 weeks before unemployment rate numbers are announced), I won’t change it since I have all the applicable polling data by that time, and won’t be getting any new information for my model.  In the same posting as my forecast, I’ll record the consensus forecast, snapshot at that point in time.  Since at that time, we all theoretically have access to the same information (ie nobody has future information), it makes sense to use the consensus forecast recorded at that time (for prediction accuracy comparisons), instead of using any future updates to the consensus forecast.

So I will use the 9.5% number for the July consensus forecast comparison, since that was the consensus at the time of my prediction, instead of the 9.6% which was recently updated, possibly due to more recent information.

Below is a graph that puts this into the unemployment trend context.

You’ll see that although the unemployment rate has been falling over the last few months, the model indicates the rate of reduction is slowing slightly. This is also evident in the polling numbers at the Gallup website.

OK, how accurate is this prediction compared, say, to consensus predictions. Let’s briefly look at my previous predictions in comparison to the monthly consensus predictions published on the Market Watch website.


First, the accuracy of the published consensus forecasts from Market Watch:

Now, my past monthly predictions:

So it looks like my prediction model, which uses Gallup polling data as the sole source of information, is very competitive with consensus estimates. In fact, my model right now appears to be slightly better, although it’s really too early to tell - we’ll keep these comparisons going to see how it all holds up over time.

It’s worth noting that even if the accuracy of my model turns out, over the long haul, to be the same accuracy as consensus estimates, there is one very distinctive and important difference between my model and consensus estimates. My model is based solely on Gallup polling data (uses no other inputs or data source), as opposed to the consensus estimates done by a group of professionals who use multiple data sources. So my model is far simpler, and if accurate, could provide a faster and easier to understand method of prediction.

Alright, until next time…

By the way, for each monthly prediction I plan to compare against the consensus forecast published on the Market Watch website. It will be interesting to see if my predictions, derived from only a single dataset (Gallup Poll Data), will be competitive with the consensus forecasts of multiple analysts with multiple sources. It turns out that last month my prediction of 9.7%  (see blog “Unemployment Improving?“) hit the actuals of 9.7% exactly. The Market Watch published consensus forecast was 9.8%, off by only 0.1 percentage points, so we were both pretty close or on the mark that time.

Next month (June forecast) will be more interesting, as the consensus is coming in right now at 9.7%, and my model predicts 9.5%. I’ve been hearing a lot of reports of how the US census is pulling back employment, and this could make the unemployment rate for June not look so good -> which goes against my prediction. BUT if that were so, the result should be reflected to some extent in the polling and therefore the model, although the polling does not capture how many people are being hired/let go. So we’ll see. Right now it’s an interesting experiment.

So I thought it would be interesting to see if we can use the voice of the common (wo)man to model and predict that good old turnaround in unemployment that we’ve been looking for. And there’s a fantastic source of information to do this - Gallup Poll interviews.

Below you’ll see some data taken from the Gallup Poll website, along with some extrapolations that I inserted. The original and updated polling dataset is at the Gallup website. According to the website, data are based off interviewing 1600 working adults and each data point in the chart represents the three day rolling average.

I used this Gallup Poll data, without the extrapolation, and was able to construct a model that generates an unemployment rate prediction (seasonally adjusted), approximately 3 weeks in advance of when the Bureau of Labor Statistics releases their number. The graph below shows you the match between the model and the actual data from the Bureau.

Although not an super tight match, it’s interesting that a reasonable correlation can be achieved with such a narrow data source - a single source of polling numbers. I am currently predicting the May unemployment number to be 9.7% based only on this model. For the curve fitting geeks out there, I will disclose that the model contains only 4 tuning parameters, and of course all the data points generated here are from a single set of parameters (no customization or adjustments by data point).

The model becomes interesting for doing “what if” scenarios based on the polling numbers. For example, I looked at the trend for the polling numbers over the last 5 months and performed a linear fit, as shown by the extrapolation in the first graph. Then, I mapped it to the model to see what it predicts. So, the conclusion, shown in the second graph, says that IF (big IF) the hiring and letting go trends continue as they have been for the last 5 months, unemployment should drop to very roughly 8.5% by November of this year.

I wanted to post that prediction here for fun, just to see how close I am on the upcoming May unemployment number, as well as the longer term extrapolated trend. I’ll be updating these original graphs with the original predictions, and overlaying the real data as it comes.

But hey, indecision is a tool for procrastination. Anybody have any advice? In the meanwhile, we’ve adopted a new little boy, and he’s keeping me quite busy - happy to say. At the same time, I’m still excited about getting this new circuits tutorial out. I think it will be really cool.

Listening to these folks, it’s clear that they don’t do it for the money. The cash is nice, but it’s not the driver. For them, the fascination, focus, and total obsession is with the idea, a tiny seed of abstract thought that can potentially grow into practical “must have” contraptions of astronomical scale. And herein lies the practical utility of abstract thinking. The abstract spawns the practical. We must have an idea to start. No idea, no practical result. And these guys will tell you that the initial idea is usually pretty crappy; it’s the constant stretch thinking and refining that transforms the starter idea from worthless to superstar. The entrepreneur has to mature his idea in the abstract before developing and executing in the practical.

The abstract stage is critical since it determines where risks will be taken as well as where all resources will be spent - essentially the farm bet. This stage determines the success or failure of the venture. Certainly, the the execution phase is also critically important, but even flawless execution is irrelevant if the seed concept is underdeveloped.

So take every opportunity you can to build your abstract muscle. We have plenty of time for practical experience; our world is filled with practical tasks. It’s the abstract thinking muscles that really atrophy if we don’t push them - and they have the biggest potential to contribute game changing discoveries.

For the math teachers out there, I found a blog of another teacher who is very focused on bringing out the practical side of math. This blog entry is filled with tips for the classroom:

Higher Abstraction Equals More Power

Here at picomonster we specialize in transforming one of the most abstract concepts, the imaginary number, into a very concrete and physical interpretation. It’s brain candy and great exercise for the synapses if you’re game - you’ll actually feel them firing away. You can try it at our Applications of Imaginary Numbers in the Real World page. Have fun!

Some common threads of advice from these speakers:

1. Win big or fail big. There is no crime in failure itself, only in failure from lack of boldness.

2. Harness brainpower and energy from others. Be a good listener and others will help to mature your idea into big opportunities.

3. Only you really understand your vision.  Don’t let the nay sayers discourage you.  Keep your eye on your passion and develop your plan to make it real.

4. Market first, technology second. Startups are tough; do them in a great market.

5. It’s not all about the idea.  You need a great team and a great market for the really big wins.

6. Total focus.  Develop your vision, create your strategy, then execute with intense laser focus.

7. In startups, speed is everything.  Beat the big fish with agility.  By comparison, they have bottomless resources, but they are slow.  Use your speed to win.

Help us get started by by sharing your experiences with us. Or you can also make suggestions for material on this site. We’d love to hear from you. If you have not yet played with our interactive imaginary number tutorial, give it a whirl at our HOME page. It’s candy for your brain!

1. Imaginary numbers provide only an intermediate and conceptual mathematical result but have no physical significance in engineering and physics applications. Circle True or False.

2. A complex number is most often used in real world applications to quantify what kind of physical variable:

3. In a vibration system with an oscillating input motion and an oscillating output motion, the imaginary part of a complex output can best be described as:

4. In a vibration system with an oscillating input motion and an oscillating output motion, the real part of a complex output can best be described as:

And another variation on the test above:

1. Imaginary numbers are far more than a mathematical abstraction; they often represent physical characteristics in engineering and physics applications. Circle True or False.

2. A complex number is most often used in real world applications to quantify what kind of physical variable:

3. In a vibration system with an oscillating input motion and an oscillating output motion, the imaginary part of a complex output can best be described as:

4. In a vibration system with an oscillating input motion and an oscillating output motion, the real part of a complex output can best be described as: