Setup is a simple Raspberry Pi 3 and camera hooked up to my wifi (< $100).

Assumptions:

This data was collected from 3/20/2017 until 8/29/2017 and has 26,000 data points, including vehicles, pedestrians, animals, go-carts, golfcarts, and motorcycles.

A breakdown of what was captured is:

## 
##     animal       bike       dark       full     gocart   golfcart 
##          5        567       3893      20266         55         82 
## motorcycle    partial pedestrian    unknown 
##         74        223        830          9

Basic summary of the speed criteria for L2R traffic (West to East) and R2L (East to West)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.59    9.61   14.69   15.13   20.02   63.13
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.74   12.53   17.90   17.89   22.90   85.89

Mean and Median do not tell the whole story, and again Speed is skewed low due to parking. Next are various histograms of the data. Each column is a count of how many cars were found to be traveling a particular speed over the time period. N is the number of data points. This is going to be the extent of the analysis until more time can be devoted to it.

The speeds break down in a slightly skewed bell curve, which is what you would expect from a natural process of cars passing by. Note there are a significant number of vehicles above the speed limit. This chart includes everything, with pedestrians and vehicles being the largest contributors.

The above chart is just vehicles where there were no significant blockages of view (e.g. street parked vehicles.) Since parking occurs on the street, very large vehicles or campers could block the view and skew the data. This is what we are working with moving forward. This chart is even more of a bell curve, which means the slower pedestrians and bikers were bringing the average speed down. Again, there are still a significant number of speeders.

This chart shows the vehicle data broken out into direction of travel. Note the significant number of East to West traffic vs West to East. 3 out of 4 cars head west, so the block is almost a one way street due to the 4100 block being one-way. Also note that the speed is skewed higher in the East to West direction, most likely form being able to see if there is on-coming traffic for a shorter distance.

Let’s see if there are different average speeds during different days of week.

Nope, doesn’t look like it, but again parking is going to lower these numbers, so maybe not the best thing to look at.

Let’s take a look at time of day (note that the earliest hours has significantly less values, so please take that into account)

Looks like there are peaks from 6:00-7:00 am, 12:00-1:00 pm,and 7:00-8:00 pm. Seems typical, and again street parking is going to lower the averages.

Since parking is lowering things, let’s do histograms for the hourly data. This is going to be a long series of plots, but very informative.

Now we can beak down when there is a lot of traffic and that it appears there are speeders every hour. The early and the late hours of these charts are skewed low because that’s when the sun rises and sets, so they are not providing good daylight for a measurements.

Now let’s break all those histograms into directions and see what that looks like

Wow. When you look at the data in this way you can really start to see the affect of the one-way on the 4100 block. In the early morning, when there are parked cars in both directions and a more evenly distributed number of vehicles going in both directions, more cars are below the speed limit. Once you get into the day you find more cars going East to West, and subsequently the speeds start to increase. It is also quite striking how the East to West drivers outnumber the West to East by almost 3 to 1 after the morning rush. It would be interesting to get data from the neighboring streets see what new traffic patterns have cropped up, especially on Alma, Leona, and Ray.