You're thinking about it wrong

Eddie Obeng from Word After Midnight covers the conceptual picture of the changing world in this TED Talk.

What can be tracked

When we talk about forecasting, there are many things that are important to consider. While we cannot consider all these aspects in regression analysis, I believe it is important to at least start a list of things to be considered as might having an effect on future events. Here are a few ideas.

• Historical Data
• Current Inventory
• Meteorological Data
• Traffic Data
• Ocean Currents
• Astrological Data
• Vibe: Positive or Negative Outlook of Participants
• Farmer's Almanac Data
• Known Future Process Data

Big Data

One way to use analytics is to go after "big data." IBM recently posted a sponsored article on Huffington Post that talks about mining RFIDs, social media and something nebulously referred to as a "quantum leap" in data collection. The problem for small business is that accessing all these areas of data can be quite expensive, especially when using RFIDs for high volume, low margin product.

Michael Schroeck, the author of the article, talks about big data in four spheres of influence: Volume, Velocity, Variety and Veracity. Volume refers to the sheer amount of stuff (data) that is being produced. Velocity refers to the speed this stuff is being created and transmitted; for instance, an hour of video is uploaded to You Tube every second. Variety means that this data is not very structured; it's in emails, tweets, videos, etc. Finally, veracity refers to the uncertainty involved with the data; simply put, you don't know what you're really getting.

So, how do we simplify this? Do we really need Ph.D.s in Statistics to get anything out of the data we have sitting in our possession already? Is it possible that the data we already have is adequate to significantly improve our lives at home and at work without sticking an electronic tag to everything? I think it is.

Start

New ideas have to start somewhere, and that somewhere is usually with more ancient ones. My concept for this blog is to start to collect ways to see the future, not literally, but in a way that acknowledges we have past history, including inventory records, sales records, weather data, astronomical records, etc. And, we also have future data. We know where the earth is going to be within the solar system for a good deal into the future. We can reasonably predict that the weather will be hot in the summer and cold in the winter for sometime to come. We understand that people do more at night when there is a full moon than a new moon. (And, I think the reasons for seasons and moon cycles and their results is self-evident to most individuals.) All that being said, there was an IBM commercial that spiked my interest. Basically, using analytics, which I presume means their SPSS platform, IBM figured out that a bakery sold more cake when it was raining out. The goal of this blog, is to find some way to sharpen historical inventory prediction methods with oceanography, meteorology and astronomy.

As far as I can tell, IBM is the only company doing anything close to this type of analysis with real world data. At this point in time, I really don't know how their doing it, but I believe the process can be replicated with some Excel, SPSS, or Minitab skill. The question becomes, "What factors or variables should be included in the analysis?" And, "Is there a broad, universal theory that can handle this data?"

Let's take a look and see what we find.