Drive Safely and Visualize Happliy - A Visualization Showcase on ATCG's AEES Solution

Author: 
Jim Sun

Summary

It is always a dilemma for small and mid-size companies to choose the right BI Solutions that meet both their cost and efficiency objectives. Constrained by their business scale, it is not possible for them to adopt large BI platforms not only because of the tremendous amount of upfront cost but also the long implementation lifecycle.

Today, we are introducing ATCG's light, rapid BI Solution called - AEES (ATCG Enterprise ERP System) and display how AEES can cure your headaches with big data analysis and data visualization. We will also digest car accident data within our system to make more sense of your massive data sets. It can also seamlessly integrate with all major BI front-end tools like:

  • SAP Lumira,
  • Predictive Analysis,
  • Qilkview,
  • Tableau etc.

Now let's get started!

Scenario

AEES in a Nutshell - A light, rapid ERP solutions for small/mid-size companies.

Link to the whitepaper.

Initiating Event

Annual traffic data has just been released by United States government traffic authorities, which is extremely important for insurance companies to monitor and respond to the market.

Background

An insurance company has just acquired a large dataset on US car crashes for 2011, however, it is a monster table with 70,000+ records, Excel hangs every time when an Analyst tries to “Pivot Table” it, yeah, you bet!

Pivot table - old school data analysis techniques.

Then the company turns to ATCG Solutions for help. ATCG BI Team loaded this big chunk of raw data into AEES, and grouped it into three major business areas (also for most organizations): HR, Finance, and Service Operations. It is a process to make the raw data “juicy” in a rapid and affordable way compared to those heavy-weight solutions like S.. and Ora… (You know what I’m talking about). But, it doesn’t stop there, as we have made our solution both Mobile and Self-Service friendly. To achieve this, the ATCG BI Team helped the company to design a semantic layer which is tailored into their needs. Last but not least, with SAP’s cutting-edge visualization tool – Lumira, now the Analyst can enjoy the BI Solution on the fly, yeah, also that grande vanilla latte on the desk.

Showcase

Source DataName of Dataset: us_car_crash_data_2011.xlsx (72310 records in total)Source: The dataset is a collection of data about people involved in car accidents with fatalities, the final injuries, alcohol/drugs tests, and other relevant data about the accident and the person. ** Original Source ** Fatality Analysis Reporting System (FARS) Encyclopedia http://www-fars.nhtsa.dot.gov


Figure 1 Snapshot of the Dataset

How The Monster Table got visualized


Figure 2 AEES Topology

Analysis Scenario


Figure 3 Generalized National Data

The Analyst starts with a bird's eye view of car crash statistics by State, since he/she is mainly working for the western coast region, the size of the bubble from California obviously draws attention. In the following graph, a trend analysis is being ran to tell the distribution over the month by gender:


Figure 4 Part of the result---Data from CA

By looking at this visualization, both Genders share the same tendency, but we do see car crash reaches the peak in July and followed by an increase till November. Thus the Analyst narrows down the time-frame to continue exploring the information.


Figure 5 Rural vs. Urban

Naturally, we will think poor road conditions and weather conditions will cause more traffic accidents. However, this visualization reveals that car crashes happened most in urban area under clear weather. In this case, you can’t trust your gut feeling. To keep digging into the dataset, Analyst adds more key elements like age, fatality into the visualization and it ends up like this beautiful bubble charts like below:


Figure 6 Severity In Terms Of Age

Eventually, the story has been told. According to our analysis, the data illustrates that in California, drivers who are 22 years old and driving in the city with good whether are most like to be involved in accidents. This piece of data is also informative to insurance companies because it helps insurance companies to know that most accidents happen during July, and our analysis can help an insurance company to better evaluate insurance pricing for different kinds of customers.

Numbers don’t lie.
Happy Visualizing!

If you would like additional information about data visualization in ATCG Solutions, don't hesitate to reach out to us by clicking the button below.