Objective

To further understand and document the impacts weather has on both traffic operations and safety.

Current Work Activities

On-going:

Analyze the data of 2013, 2014, 2015, and 2017 data to make our conclusions more reliable and accurate. The key point of this stage is to use the Microsoft Azure to preprocess the data, as the raw data size is up to the TB.

1) Data Collection:

• Target: collect Wavetronix, INRIX, and weather data in 2013, 2014, 2015, and 2017
• Progress: All the data are collected, except the INRIX data in 2013 and 2014. Data is in downloading.

2) Data Uploading to Microsoft Azure:

• Target: uploading the raw data to Azure. It is non-trivial due to the big data size.
• Progress: since it is very slow to upload big data from our desktop, we will send the data to th Azure Company to upload them as soon as they are ready.

3) Data Pre-process and Analysis:

• The tasks of multiyear data analysis are summarized in Table 2.

Next Steps (if needed)

Planned tasks of multiyear data pre-process and analysis:

1. Data pre-process
2. Evaluate the performance consistency of Wavetronix sensors under various weather
3. Extreme weather analysis s weather
4. Analysis of the dynamic change of weather effects on travel speeds

Completed Work

Preliminary Analysis Work:

Beginning point for quantifying the weather effects on travel speeds

1. Identify the characteristics of weather in space and time
2. Quantify the weather effects for each sensor with the linear regression model: snow, cold rain, and warm rain.
3. Use the K-means and latent class clustering method to classify sensors into three categories by weather effects

Deep analysis to extreme sensors

1. Identify extreme sensors with the linear regression results: focus on analyzing the sensors where snow and rain show significantly positive effects
2. Illustrate the causal factors leading them to be extreme: geometric and traffic
3. Compare the performance of extreme sensors under normal and adverse weather to figure out their correlations
4. Develop the algorithm of determining whether removing extreme sensors

Deep analysis to extreme weather events

1. Identify the extreme weather events by checking the daily CDF of travel speeds
2. Develop the thresholds of determining extreme weather events
3. Identify the influencing range and degree of extreme weather events by plotting the distribution of travel speeds over space and time
4. Explore solutions to relieve the influences of extreme weather events

Deep analysis to weather effects on traffic of Des Moines

1. Analyze the differences of weather effects on the bidirectional traffic of the ring freeway of Des Moines with the linear mixed effects model
2. Identify the weather effects on the traffic of morning peak and afternoon peak
3. Figure out the differences of weather effects on traffic in Des Moines and surrounding suburban areas
4. Explore the joint influences of adverse weather and traffic incidents on travel speeds