Public Data on DataBaselines

Georgia Exports – On Track for Topping $40 Billion in 2018

Tue, 15 Jan 2019 00:00:00 +0000

Export goods originating from Georgia reached an all-time high-mark of $39.4 billion in 2014. As a state, Georgia has been trying to exceed that banner year ever since. Recent talk of tariffs and revisions to NAFTA have caused many Georgians to wonder what effect this may have on our economy. We will have to wait until February 2019 for the US Census Bureau to publish complete 2018 export totals. But currently available data can help us understand the role that exports have in Georgia’s economy and what countries dominate as our trade partners.

In 2017, Georgia’s export economy was the 12th largest of the 50 states. Over 15 years’ time, the dollar-value of goods originating from Georgia and sold abroad has grown 2.5 times. In 2002 exports totaled $14.4 billion. By 2017 they had grown to $37.2 billion.

Georgia Exports - Primer on Using Census Data

Fri, 11 Jan 2019 00:00:00 +0000

Our company is interested in exporting our products. What data sources are available and how can we evaluate our prospects?

The U.S. Census Bureau has a wealth of information for evaluating both domestic and international markets. In this series of blog articles, I’ve documented the methods for retrieving Census export data and analyzing it in Power BI. Topics include: the product classifications used by the Census, the mechanics for retrieval and how Power BI and the Tabular data model can make quick work of a 5 million row data set.

Viewed in aggregate, you can see that exports have an important and growing role in our state’s economy (graph below.) That data really comes alive when you can drill down and explore data trends in your own industry.

Change over Time Analysis in Power BI

Wed, 02 Jan 2019 00:00:00 +0000

In an earlier post I showed you a Power BI interface for summarizing Georgia export data from the US Census Bureau. If your primary interest is year totals, that model is useful for understanding our State’s export economy.

But there are some questions that particular Power BI model does not answer. In particular, it does not address change over time questions like:

What product categories have the greatest gain or loss?
What countries were responsible for those gains and losses?

This blog article and an accompanying video explain how the updated version of our Power BI model address these questions. The video takes you step-by-step through the system. The paragraphs and images below are a quick orientation for you. Note: clicking the images will increase their size.

Trade Data via the Census API and SSIS

Wed, 05 Dec 2018 00:00:00 +0000

This article continues an exploration of the Census Trade data sets, but with an eye for doing larger “bulk-loading” of data the Census makes available. Specifically, we will:

Retrieve export data from the US Census by using their Application Programming Interface (API),
Import the API-generated data file into a database using Microsoft SQL Server Integration Services.

In a previous article, I described a method of generating export data using the Census USA Trade Online website. That method is great for initial discovery and extraction of one-time data. But if you are ingesting Census data on a recurring schedule - you need a process you can repeat easily and consistently. The Census API is useful for this purpose.

Georgia Exports in Power BI

Fri, 30 Nov 2018 00:00:00 +0000

Describing Export Statistics using Power BI
#

You may already know that Exports from Georgia contribute between $35-40 billion to our state economy each year. But have you ever wondered – What do these Georgia products look like?

This blog article and attached video is an effort to answer that question. It uses a free tool from Microsoft called Power BI to slice and dice the detailed monthly export data that is available from the US Census.

What Does the Power BI Tool Show?
#

I’ve previously retrieved this Census data for 2013-2018 and ended up with a 2.4 million row data set in SQL Server. While it is all well and good to have the raw data - the trick is putting that data in a format end users can easily navigate and learn from. On this score, PowerBI delivers.

Georgia Exports, A Tutorial for USA Trade Online

Thu, 13 Sep 2018 00:00:00 +0000

Trade exports accounted for $37 billion of Georgia’s economy in 2017. That translates to over 180,000 jobs in our state. With so many revisions to tariffs and NAFTA taking place, you may wonder: “What effect will these changes have on our State’s economy?”

Starting with this blog article – I’d like to share what I’ve learned about the subject. Together, we’ll evaluate the most authoritative and detailed source of US import/export data that is available: the US Census Trade data. In this article I’ll show you how to extract data from the Census using USA Trade Online. This is an easy-to-use and powerful method of retrieving trade-related data. In later articles I’ll address:

Using R to Evaluate College Scorecard

Fri, 03 Nov 2017 00:00:00 +0000

Our company acquired a data file containing over 15,000 rows and 300 columns. We are trying to identify patterns in the data. Where do we begin evaluating such a large dataset? Would Using R be helpful?

Let me answer that question by taking a look at data from the U.S. Department of Education known as the College Scorecard. This organization compiles a wide variety of measures for all colleges and universities in the United States. They track over 300 measures such as:

Average SAT scores for incoming students,
Percentage of students whose parents completed high school, and
Whether the Institution is private or publicly owned.

I originally evaluated this data as part of a capstone course in the Microsoft Professional Program for Data Science. In that class, the overall goal was to identify a subset of these 300 measures that were most useful in predicting a single measure – the average income of graduates. Admittedly, it’s a bit overwhelming to know where to start resolving this sort of problem. You have intuitive hunches that SAT scores or faculty salary might be correlated to the future income of graduates. But how do you systematically evaluate a data set this large?