[MUSIC] In the previous lectures, we have discussed four disciplines related to spatial data science -GIS, spatial DBMS, big data system, and data analytics, and open source software for each discipline, and then system configuration dependent on characteristics of given problems. Now, we'll discuss materials - data issues. In this lecture, we'll discuss what spatial data are and the examples, then introduce spatial big data, its definition and the examples. In the end, I'll give you a brief overview of the value of spatial big data. In other words, what we can do better with spatial big data than with only spatial data. [SOUND] Spatial data can be defined as data which has location information on coordinate system, and which can be mapped. They generally accompany attribute data, which describe name, condition, and other information. Spatial data is also known as Geospatial Data, Geodata or GIS data. Spatial data is represented by way of two different types of data models, vector data model or raster data model. You'll learn about it in the third week in more detail. There are many examples of spatial data, such as parcel boundary, road network, river network, point of interest, we call it POI, land use, land cover data, digital elevation model, which depicts surface of the Earth. Orthophotos, which are photos as well as maps, some 3D maps, which are recently available in many applications. Spatial data examples, in black letter, are vector types, now you are looking at, in green letters, are raster types. Now let's take a look at real spatial data. On the first row parcel map, Road network in Korea, and subway stations in New York, they are vector format. On the second row National Land Corps data, it is called NLCD of State of Washington and Oregon, produced by USGS. Landsat-7 satellite image, digital elevation model, DEM, of part of Louisiana, they are all raster format. On the third row, 3D indoor map, 3D point cloud, 3D building models are given. [SOUND] Now let us discuss about spatial big data. Spatial big data can be defined as spatial data with big data characteristics, or the opposite way around, big data with spatial context. What is big data? Big data are defined as datasets which are so large that traditional data processing are inadequate to deal with. They have three characteristics of volume, velocity and variety. There are many spatial datasets with such characteristics. You're looking at examples. Floating population data present the number of person, and their sex, and their age ranges with respect to given cells as spatial unit per every minute or every hour as temporal unit. Vehicle trajectory, which is well-know spatial big data, for example, taxi trajectory in NYC, New York City. Card transaction data with location, such as credit card and public transit card, SNS data with location are also good examples, like Facebook and Twitter present the geo-tagging to message or activities. Furthermore, there are even location-based SNS's, such as Foursquare. A collection of sensor data with location also produce spatial big data, right? Digital tachograph, DTG for truck fleet operation monitoring is a good example. As a different category, there are spatial big data in image format, in other words, raster format. Geo-tagging photography on Flickr and weather data for a single country or the world, they're examples of special big data in raster format. From the perspective of spatial data analysis, earth-observing satellite images should be noted. There are quite a few satellite imaging systems, which generate a humongous size of spatial data every day. For example, Landsat 7 system can theoretically downlink 17 terabyte of satellite image data per hour. Now let's take a look at spatial big data examples. On the first row, floating population data, tax trajectory, and public transit card transactions in Seoul, Korea are given. On the second row, average temperature of the US for single day as weather data, and check in page of Facebook, which produce geo-tagged social data. And on the third row, geo-tagged photos, people's trajectory in terms of semantic locations. [SOUND] There are many application fields of spatial data science, particularly with spatial big data. Now you are looking at a list of application fields. They are eco-routing for minimization of fuel consumption, urban traffic analysis for optimization of transportation service. Real-time city operation for disaster response and mitigation. Geo-segmentation for marketing with floating population data. Let's say, for example, questions like where is the best place for clothing shop for young ladies? Other spatial business intelligence and geo-political analysis with geo-tagged SNS data, they are good examples. For natural science, climate change study with satellite image analysis would be a good example of spatial big data. Another example I have is healthcare application, and take a look at it in more detail. The given analyses are service accessibility to medical emergency rooms with two different population data, population of administrative district, on the left, and floating population from cellular phone user's movement on the right. The first analysis is conventional, and the second becomes feasible only after spatial big data available. Which one do you think is more reasonable and accurate than the other? The conventional approach on the left came up with a biased conclusion due to consideration of only residential population. As a result, center of business district, CBD, could become the places of best service accessibility, due to very little residential population. Does it make sense? Do you think we need emergency room or emergency service only when we stay at home, of course not. The analysis with floating population would give a much more accurate and insightful analysis regarding the given problem. [SOUND] In this lecture, you studied on spatial data and spatial big data as input materials for spatial data science. I believe, spatial data science gains more and more importance and significance in many application fields, after we have spatial big data available, which can present better 'insight' as well as more 'accurate' analysis. This is the end of the second week, in which we have discussed related disciplines to spatial data science, softwares, solution framework, and datasets. Hopefully you have enjoyed it, next week we'll discuss each discipline of spatial data science in more detail. See you then.