Chief, Statistical Methods Branch
Agricultural analysis leads to vital economic decisions
USDA's NASS collects, summarizes hundreds of data series about the nation's crops, livestock
Think of it as the database for the US Department of Agriculture. Every day, economists, insurers, equipment manufacturers, transportation/logistics experts, textile firms, food producers, regulators, traders, investors, speculators and government leaders rely on the USDA's National Agricultural Statistics Service (NASS) for a broad range of information, analyses, forecasts, and other vital data that help them gain a clearer picture of the country's agricultural status.
With a firm commitment to timely, accurate and useful information, NASS conducts hundreds of surveys annually and prepares reports and analyses that cover virtually every aspect of US agriculture – including production and supplies of food and fiber, prices paid/received by farmers, farm labor and wages, farm finances, chemical usage, and changes in producer demographics.
But there are also hundreds of other sub-estimates, surveys, and data series for dozens of livestock species. For instance, there are semi-annual and monthly cattle reports covering milk cows and feed cows, breeding inventory and marketing inventory. Then there are the estimates for more than 120 different crops, such as cotton, oranges and wheat. A typical cattle survey might see 40,000 ranchers responding to a 20-question survey or answering questions in a 15-minute phone interview.
Given the data intensity of the challenge, the analytical software foundation must deliver exceptional levels of performance and sophisticated functionality. According to Dave Aune, Chief of the Statistical Methods Branch for NASS, SAS® tools have been the basis of the agency's data aggregation and statistical analyses since the mid-1970s.
"We actually may have been among the first SAS customers nearly 35 years ago," he said, "and there could be only one reason why SAS is still at the heart of our statistical processes: It works exceptionally well."
SAS® helps pull it all together
One of the biggest challenges that NASS faces is simply pulling together all of the data. The costs dictate the modes of data collection used, depending on the particular need. Mail-in surveys are least expensive, but necessitate turnaround times that can be challenging for monthly data – sometimes as little as 10 days. Face-to-face interviews by field researchers going out to farms and ranches almost always yield higher-quality data, but are expensive and time-consuming. As a result, cost-effective and time-efficient telephone interviews account for about 75 percent of NASS' data collection.
SAS not only pulls the information together for us, it uncovers the issues we need to address before we release the official USDA estimates to our various consuming audiences. This information is, of course, extremely sensitive and important – it's often a mission-critical indicator for a critical segment of the nation's economy, so we can't allow any errors to slip through.
"We have 45 field offices conducting these surveys," explained Mark Apodaca, Head of the Commodity Section in Statistical Methods Branch. "We capture that data in Sybase and Redbrick databases. From there, we've built interactive analysis tools in SAS that pull survey responses for review and calculate expansions and ratios.
"This is all collected and performed at the state level, which has access to these prebuilt SAS routines so they can turn things around quickly. We built these routines in the 1990s, and they are still in use. Our staff in the field offices are subject experts in agriculture, not math. They're not required to be SAS experts. Instead, they can point-and-click to view questionnaire-level data quickly and easily to review the responses, make any edits, and prepare for the submission."
Spotting – and correcting – the anomalies
"Here at our central office, we're also heavy users of SAS on the mainframe," said Aune. "Before we consolidate the data from our field offices, we spend a fair bit of time reviewing it to validate its accuracy and finding anomalies. Instead of spending time detecting problems, we can spend more time researching the underlying issues. Once we're satisfied that the data are clean and correct, we aggregate the data and build a national summary. SAS not only pulls the information together for us, it uncovers the issues we need to address before we release the official USDA estimates to our various consuming audiences. This information is, of course, extremely sensitive and important – it's often a mission-critical indicator for a critical segment of the nation's economy, so we can't allow any errors to slip through."
"SAS is just synonymous with the work we're doing here," said Apodaca. "It's part of the skill set we look for when we're hiring. In fact, one small reason why SAS is such a strong presence here is because most of us were immersed in using SAS in college. Today, we're intense users creating summaries, ad hoc analyses, anomaly exploration, data massaging and more. It's an indispensable part of our daily work here."
Capture, edit, summarize and report data for hundreds of data series about crops and livestock across the US; detect anomalies lurking within the data.
With SAS, NASS can more easily aggregate, analyze and review vast volumes of sensitive and critically important data that forms the basis of economic forecasts and far-reaching business and government decisions.