NWGIS 2018 Conference

Do you ever struggle to find the exact GIS data that you want?  Do you get frustrated that available datasets aren’t granular enough or are redacted in a way that isn’t actionable?  We live in the Big Data revolution, so why should we give up if we don’t find the exact GIS data that we are looking for?  In this session, we will explore the idea that to get the GIS data you want, one dataset, you may need to employ and combine multiple datasets.  To demonstrate this, we have a case study on enhancing international wealth and demographic data, although this approach can apply to any GIS dataset and to any GIS industry.  Case Study - Enhancing International Wealth and Demographic Data: For many reasons, US data about demographics is easy to find, even on a micro-geographic basis, but in many other countries around the world, finding micro-geographic data about demographics is hard. In the US, we are fortunate to have the Census Department, which offers extensive data about things like the number of people of various age groups. This data is available on a micro-geographic basis, which technically is called a “block group” – polygons containing roughly 1,500 people. In markets outside the US, getting demographic data totally depends on the country. In China, demographic information only comes in very large geographies, including on average 500,000 people. That means that US data is, on average, over 30x more granular than Chinese demographic data! Wealth data in China is essentially non-existent on anything but a city-wide basis. So how do you get the data you want? In this session, attendees will learn the principles we use to turn big piles of data into micro-geographic information: examples of data that enhance demographics and wealth include road networks, urban greenery data, real estate data and building volume data. Examples of principles include materiality, source quality, structured/unstructured data and strategic intent. The moral of the story we are telling is that the quality of a single dataset can almost always be enhanced. So, just because the data that you want doesn’t exist, doesn’t mean that you can’t estimate it.