地理空間データ：ベクター、ラスター、解釈

[0:04] Hi. My name's Todd Bacastow. Welcome to the second lecture of lesson two. We're going to talk about geo data, geospatial data, and particular geospatial data as symbols. Data are symbols that represent measurements of phenomena. Number of ways we can do that. We can do it as pictures. We can do it as words. We can do it as numbers. These measurements are characteristics at a location. They represent something it's going on.

[0:38] Why do we use symbols and make data? That's a good question. Because it allows us to do things easier. We can store it and once we have it in a symbol, but also we can then begin to process that. We can do other things with it if we want to study it. It allows us to look at certain characteristics in certain ways at certain times. How accurate the data are actually depended on how we collect the data, when we collect the data, where we collect the data, and what aspects we collect.

[1:11] Sort of important things. Data can be viewed as a commodity today. Big data's the buzz word you hear all the time. Being collected all the time by entities, Google's collecting data on you. The value of data's not necessary lost however when it's used. For that matter, being old it still has value, in fact in some cases it may have more value because you can see how things have changed if you collect current data. We mentioned before though that data can also be transformed into information.

[1:42] And then that information can be transformed again back into data. So data's sort of an interesting thing. Data are samples. What do we mean by that? The Earth is just too big and detailed to collect all, all the information, all the data. So we go out with a scheme in order to sample the Earth, that allows us to represent the Earth, we hope in an accurate way with a lesser number of points. There are two primary schemes for sampling the Earth.

[2:15] Vector and raster are the two schemes. The vector approach or strategy involves sampling along lines. It may be along the length of a road. It could also be around the perimeter of a building. We collect information at points or locations, and we connect these points together with a line, which will give us a linear feature, a road. If we close the line, it will give us a polygon which could represent a building. Vector data model is how surveyors survey a property boundary.

[2:53] The vector strategy is well suited for mapping entities that are, have clearly defined edges. Highways, personal boundaries, things that we know where the edge is. There's another alternative way of plucking this information, it's called raster. The raster approach involves sampling attributes at a fixed interval. Each sample represents one cell in a checkerboard like fashion. The raster strategy is a good choice for representing phenomena that lack clear boundaries, like elevation.

[3:26] Where's the edge of an elevation? The important thing about raster is digital airborne and satellite imagery today are largely raster data collected by running a scanner across the Earth's surface. This is a pixel by pixel, row by row collection. What's the bottom line? Both raster and vector approaches accomplish the same thing. We are collecting samples about the Earth's surface.

[3:59] They allow us to record and to characterize the Earth's features with a limited number of samples since we can't sample it all. How do we distinguish between these two? Let me give you two analogies. The vector approach is like a stained glass window where we have lot of shards of glass that are brought together to cover the entire surface. The raster approach is like a tile floor. We have very regular pieces of tile in a uniform pattern to cover the entire area.

[4:34] However, if you think about what we've done with both vector and raster approaches, we sampled. We have looked at the Earth's surface and we've dissassembled it. In order to use this data, we need to reassemble it the way we want to. In the process of dissassembling the Earth, we've also simplified. We have to put some understanding back into the data. So data without understanding really just gives us meaningless symbols. We do collect metadata about our data.

[5:04] But that metadata, by and large, is technical details of the data. File size, file structure, et cetera. What is really necessary when you're in the GEOINT realm is to put knowledge back in there. We term this techne. It is the human craft of bringing knowledge back into the data. It is the process of adding understanding back into the data when we fit data into a frame. And I'll talk much more about this frame later on.

[5:38] Techne is term derived from Greek. It means craftsmanship, craft, or an art. It could be also described as data know how. But really it's much more than that. It's wisdom. It's the quality of having experience, knowledge, and good judgment about the data. Much of this wisdom is acquired through experience. The analyst uses the wisdom when handling the data in process to form judgments about a place. Data wisdom is difficult to communicate.

[6:12] It's very difficult to write your wisdom down in words. This is often termed tacit knowledge. Tacit knowledge is learned through experience. If you think about a stonemason, you can read a book about stone masonry. But you really will never become a mason until you apprentice with someone that shows you the art. It's a craftsmanship. It's gained through observation, imitation, and a lot of practice. What I'm suggesting is, this tacit knowledge with GEOINT data when we put it back together, that knowledge is gained through observation imitation and practice.

[6:54] As we discussed earlier, the analyst's craft of fitting the GEOINT data to a frame includes technique. It is that wisdom that brings meaning to the symbols of a place. However there are typical understandings of place we may have. There are a lot of conceptual models about how places are organized. This will help us build a container or skeleton that we can put the data in. So we can give it some understanding. Give it some meaning. Knowing the typical geospatial organization of a place combined with the data will allow us to frame something.

[7:30] These frames allow us to make judgements or educated guesses about some qualities of the area. If you think about it, we do it everyday. If I'm driving my car and I'm on a certain road, and I see a certain situation in terms of both the physical arrangement of cars in front of me, maybe a sense that I have that an accident's about to happen. I'll potentially take some evasive action. That's from experience. It's from experience of driving my car. Theories often form as basis for spatial organization.

[8:03] They'll allow us to frame things. These theories help us build mental models, and they're relationships between things we might not recognize without the theory behind it. Geography has a number of great theories. Just to name a very few, examples of important theories to help us build models are the Gravity Model, Central Place Theory, Weber's Model of Industrial Location, Von Thunen's Agriculture Model, and the Core Periphery Model.

[8:33] There are many more. They help us organize what we see. They give us a framework. They give us a skeleton. I'm showing you a picture of the Central Place Theory. If you've ever flown across the Santo plains of Germany. If you looked out the window of your aircraft you might see the cities and towns configured somewhere to that. So a frame provides us a cognitive skeleton. It helps us to arrange the data in a meaningful way. A frame can take a number of different forms.

[9:06] It can be a narrative. That narrative can be a chronology of events and causal relationships between them. A frame can also be a map, if you think about the map, it has distances, directions, connections. It helps us organize things. It explains, it gives us a story of what's happening. Frames can be thought of as models. Like a map sort of as a model. Can also be thought of as generalizations that are at work, like a map. Frames however are not theories.

[9:38] But a frame may be influenced by a theory. For example the Central Place Theory may influence how we frame something. Framing is normal. It's an everyday activity essential for geographic problem solving. In fact maps are a frame. GEOINT data and frames help us make sense of what we see. Let me define sense making. Sense making is the ability to understand ambiguous situations. Three primary outcomes of this understanding.

[10:13] Identify patterns, describe patterns, and predict future patterns. In sensemaking, joint data are an essential part of a mental dialogue to explain what the data shows in the context of how we frame it. Here GEOINT's making the bulbs continuous cognitive work to understand the relationship between the data, and the place and events in our frame. Let me summarize what we've learned in this lecture. GEOINT data represents a place while conveying a sense of physical and human qualities.

[10:47] In addition, GEOINT data requires understanding about a place's spatial organization to make sense of what we see. Knowing the spatial organization allows us to frame human activities to understand the processes which create these patterns. The depiction of this GEOINT data with respect to a known model is a frame. Frame guides to search for additional of data. The frame ultimately provides us a story or a map to account for what we see. Thank you very much.

[11:17] I appreciate you joining me.