[MUSIC] Hello again everyone. In this lecture, I'm going to go in depth into how vector data is structured. We'll talk about how vector features are created, the object model of vector data, and precision of vector calculations. This knowledge will be foundational to your concrete work with future data in this class and to the ones that follow. As a refresher, since it's the beginning of this class, let's start from the very beginning with vector data. Vector data is one of the two core common ways we classify data in GIS systems, with the other being raster data. We often conceptualize vector data as points, lines, and areas or polygons on a map. This makes sense because if we look at the name vector, it refers to something that has direction and distance from a location. In the vector data model, remember that we start with a coordinate system and build everything else on that. Locations are given directly as coordinates in that coordinate system with the most basic locations just being a single point given as a pair of x and y-coordinates. This location has an associated attribute record through which we can describe the features of the object the point represents. From here, we can build the rest of the vector data model. If we connect multiple points together, we can make the line data type, built of point locations with information on how those points are connected to form a line. The line data type has attribute records too, since it's still a vector data format. But instead of that attribute table record referring to a point, it refers to the whole line. Now with polygons, we can do the same thing, making lines to form a boundary and our actuates refer to the whole area contained in the boundary. To make all of this work, we need to think in terms of generalization. How can we take specific, real things and represent them as some general location and set of attributes? How can we group all objects together? What attributes do they share? To do this, we use a data model we often call the object data model. With objects, we're conceptualizing real world items as some approximation of that item as described by a set of attributes we define. These can be concrete things like human, or abstract things like ideas. Either way, for a given objective, there might be things that we need to know about humans. Maybe they are participants in a study, so we assign attributes for these variables of interest about a person and fill them in for each record. GIS data is no different but one of those attributes is the location information. Let's take a look at these buildings for a moment. If we represent these buildings as objects in a GIS system, we can start defining a common set of attributes for them. This building is 205 meters tall, this other one is 180 meters and this one is 176 meters. The primary building material for the three of them is steel but this bottom one is mostly brick. The facade of this tall building is glass and these two are concrete and glass and again, this one over here is brick. We can also record the number of inhabitants, the hours it's open, etc. And we could represent these buildings as points or even as incredibly sophisticated three-dimensional floor plans. Regardless, we have an object where we've generalized the qualities of something so that we can compare it to others, or just work with it as a concept on a computer. When we have a group of these objects, we can organize them into a data table with the field or column names coming from each quality, height, primary building material, etc. Again, with the location information as a part of that record, too. The data table lets us group the data, providing a common way to access data representing the same quality so that we can view, update, query and sort it. This is an incredibly useful tool to have. Some of you are probably thinking, this is obvious and others are probably thinking, this is still confusing. The takeaway message here is that records in a feature class represent some concept of an object or idea in the real world. Each feature class is a collection of many of those objects and then we represent the locations of these objects in some generalized form too as points, lines or polygons on a map. Now that we have a good sense for vector data structure, let's talk about how tools work with it. There's an old saying that's quickly becoming outdated that says raster is faster but vector is corrector. The main point of the statement being that raster calculations often finish more quickly and while vector data is more complex, it's often more accurate. As with most generalizations, this phrase has its limits. But what it does point us to is the idea that vector calculations are precise but harder to do. Think about it this way. For a computer to figure out if two polygons overlap each other, it has to determine that based on just a few points around the edges. It then has to construct a concept of the space for what's inside each, and determine if any of that space is shared. That's complex math, especially if those two polygons were in different coordinate systems. But it can be done with precision and vector often makes the most sense for data that we want to generalize but where it's easy to generalize properties of large areas into the same feature. Similarly, re-projection of data from one coordinate system to another in the vector data model is comparatively precise. We'll go through why it's harder for rasters a bit later in the course but this property makes vector data a reliable way to store features if you want to capture precise locations for point observations, edges or boundaries. Similarly, since many modifications of the data don't effect the attributes, those attributes stay reliable so long as they are appropriately joined into the data. The last thing to mention here is that vector and raster data use different tools in most cases. Some tools work on both but because the operations are so different, they frequently require different algorithms and different tools. This is confusing at first but it gets easier as you become accustomed to the work flows for different data types. At some point, you'll stop thinking of raster data in such a way that it seems appropriate to use the same tools as you would on a vector data set, and vice versa. Let's stop here. To sum up, vector data follows an object model, where we select attributes of items of interest, and we gather them to create a conceptualization called an object. When we collect them together, we get data tables, and when we add locations, we get feature classes. These locations are constructed of points or points chained together into lines or lines chained together to make polygons. Ultimately, it's a data model that provides good precision for our geospatial calculations. In the next lecture, we'll demonstrate some of these qualities in ArcMap to help cement these concepts. If this is all confusing for now, don't worry, you're not behind. It'll become clearer over time. See you next time.