Sometimes when you're looking at a dataset, you might look at the attributes, so, like the field headings and things like that, and it may not be really obvious to you what it is that you're looking at. This is something that I think you need to get used to when you're working with a lot of GIS data, is just getting data from someone and saying, "Okay, what do we got? What's in here? What is all this stuff mean? " So, here for example, if we're looking at this ONrte file from DMTI, we have an FID column, a SHAPE column, STREET, FROMLEFT, TOLEFT, FROMRIGHT, TORIGHT, PREDIR, PRETYPE, STRERTNAME, SUFTYPE, it goes on, and on, and on, and on, and on. So, what does all this mean? What is the thing we can use to interpret or understand what all of these headings mean, how they were derived, how they came up with this? To do that, we would use a data dictionary. So, a data dictionary is a type of metadata, if you want to think of it that way, that's a detailed description of the data contents of a database, and with particular attention paid to the explanations of categories. So, this is not something that's necessarily automatically included say in our catalog, this may be a supplemental document or data file or something like that, that explains what all of those field headings mean and so on. So, let's have a quick look at an example. So, for this particular dataset, there's a PDF that came with it from DMTI spatial, and they don't actually call it a data dictionary on the cover, you don't see it over here anywhere, it's just documentation but inside and the table of contents, you'll see that there is a whole section called data dictionary, so what is that? You'll see that they have this for all of the map layers that they're producing as part of this larger product that they sell. If we go to the data dictionary section for the same file we've been looking at the roads data, you'll see that they have a little screenshot of what it looks like, they have a list of the atributes fields that goes on through here, and then they have what's the content of each of those, and just as an example, I will go through the whole thing, is you can see that they have things like the explanations for all of those field headings I was just looking at, like the FROMLEFT, the TOLEFT, PREDIR. For example, PREDIR means prefix direction, now that makes sense, component of the street title. So, if you're you say, in New York and its west fifth street, the W there is the prefix direction. Not to get into a whole attention on that part of it, but that's what a data dictionary is good for is saying, "Oh, so that's what that field heading means." Then you can decide whether that's useful for you. It goes on, for example, one thing that's important is that they have this thing called SPD_KM which estimated speed limiting kilometers per hour, and you think, "Okay, so they went and they found the speed limit for every road, in this case, in Canada, and they included on the database." Well, actually not quite, if you look at a little more closely, let's see what we find. The description for that field says the estimated speed limits are derived using the Cardo value of a road segment and the population density in the vicinity of the road segment. Populated areas have a population density of at least 100 persons per square kilometer, sparsely populated areas have a population density of less than 100 persons per square kilometer. Okay, so, what this means is, in other words, that those are not speed limits that have been collected from driving down every road and looking at the signs or talking to the police or the city or whatever, they've been estimated based on what they think are the population densities. So, in other words they've just said, "Oh, if it's a higher population density, that means the road would be slower, " and you think, "Okay, why is he going on about this?" Well, I'll give you an example. Is that I was helping to supervise a graduate student here at the University of Toronto who was doing a study on the response times of ambulances to trauma incidence, and he was using this data to help estimate the travel times for those ambulances. It was my job, this was somebody in medicine, so I certainly don't know a lot of what medicine but I do know a little bit about GIS data, and I was there to point out these things and say, "Well, if you look at how that was generated, that's really kind of a rough estimate." I mean, I'm not saying they should never be used or DMTI did anything wrong, but if you're trying to improve ambulance response times, just make sure that you have something in your study that says, these are estimates, there's assumptions that are being made and you wouldn't want it to turn into a situation where a poor decision was made based on data that didn't have the level of specificity that you thought it had. Now, in that particular case, that was all the data that was available and it certainly was better than nothing and gave them a good estimate and they can certainly look at relative differences quite well and get a pretty good idea, but I just felt it's important to always say, well, just make sure that you know what you're actually using and that you have a good idea what that data really represents.
