Merken

Creating Map Style & Visibility Rules from Statistics

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
and that but it was no more room so you know so on and on and on and on the Thieu hand uh and and and and uh it but it was no more room so you know so on and so on and so and so on and so on but you Thieu and a and a and and that that thank you for coming to this presentation in this is not a serious map by the way so unless you live and die in box world unboxed it's something think anybody does not in the way of and also thinking thank you for the last 2 conference coming to a presentation that contains the word statistics in the title of and I tried to come up with a synopsis of the statement and then I whatever had beer I thought if my title and put you to sleep this would but basically and I have to admit that I was so it's a somewhat to about forming this like you come up at different angles to try to follow about this subject matter whether to focus on specific apps like map server which is where my experiences and Cordelia OK it from yeah of the focus of the war it will think because that the idea myself out of the weight of the where the started was or my background and what I want to study this is I usually work for my maps and map server with POS GIS datasets and I found myself spending on more time than I wish on trial and error optically around hostiles development on around how they relate to class breaks how they relate to stand on the scale of ranges where the scale of max and min denominations are and so forth for different classes and and for different layers and so forth and can also I work with the data set it's common in the United States the U Census Tiger dataset which 1 is and
all USA resource which is comprehensive but error-prone and in particular I want my conditions I run into lot where if the classification system is either it is plain poor it's not a it's either not applied well like as and all blanks and into instance inconsistencies are it's so it's not really directed at the the task of mapping as well as I would like stuff on it to
turn them and it also this even know
what I'm talking about what I've been studying is somewhat similar to the medical mapping it's not really what I'm after I'm after more styling for a basic space map and the saw not specifically telling a particular story of like go on I stole this from some Google Images page the whatever whatever thing they're doing here is very specific and that's not really what I mean when and England really have a total the best base-map story of so they're important features of our uh emphasized over a not so important features and so forth and and and so on I think this is just a the C. young lots of about the classification schemes that are already in the data have mostly mention these points already that I that you may find that the sum of the copy of classes you want for mapping may come from a multitude of columns in the table and the various expressions of their or of I in the the Our outcome cases where the party scheme just doesn't give me a clue about how I would want to prioritize and on and so I found in in research in this is that the virtually 2 sides of the equation 1 side is the setting sun basically what are you saying like and again I'm usually and working in map server I'll show you a minute down of some of the work I do in there and so but also like that I'm basically trying to focus on some deletes and troubling problems where in my work with trying to set up a map far and for the base map portion of finding that you know my my solutions don't fit everything perfectly of come into conditions where the condition of crowding our how stacked layers aware of set the scale breaks out classifier was symbols what fonts would let you know everything using they apply mostly in and find the breakdown in certain spots and if I try to analyze the spots you'll find it's a very and it's a sometimes it's a subtle on condition it's hard to describe even harder to describe a solution for and and then there's a series of what I called hazards are limitations to to I trying to set up in the scales in such I for example you're you're dataset you may have of constraints on your data where you cannot altered are you may have a mapping scheme that you accord follow where the number scale breaks and so forth are already set and so you can't the doctor that but some of and I the but in 1 of the things I've found with the tiger dataset and actively working with rogue data is the original roads are partitioned not really for the purpose of mapping but for addressing maintenance stepped in order for them to collect their US census records and to provide the foundation for the demographic and mapping they've partition the streets at intersections with each other and with the other features and which is nice for recordkeeping but it's not so great format is and you get more segments and you really wish now they also model in recent releases Tiger they they model from they model the road more on a map from way although it's it's got its own party label partition at US county boundaries and other things that I might not wish to see in some of my map views and and I am not 1 of things I well I did is briefly went through some of the existing projects outside of map server to see where's there are some clues about how to use the form and statistics about the data in order to classify a captured a string here from GIS where in order to set of colors that say I want a it's called a graduated where you can assign of several different schemes to class breaking you can pick the number of class breaks and you can pick 1 of the smaller the color schemes this case there was a green but I guess it's a green saturation point scale and this particular 1 is called natural breaks were it get it and tries to find the the separation is the inherent in the that part of the data set into 2 in order to them of In order to make the groupings that you choose to follow the natural groupings that are inherent in the data and on and then no trouble so looked I have not studied I'm not a person so I have not extensively studied user but I did find a few things in the documentation that I thought I would give some of help in trying to assign the data values to style values and map server has a feature that I'm not used to for this that it's not will documented there's where you can specify color range and and what the data values are that that are assigned to that and what what data column that is and and so in order to in order to these are the basic types of styles that style and layering another of settings that I have to work with in generic terms it's C the order in which you layer and for purposes of of map called I think 1 of the things that I you need to analyze is you want particularly happens when a particularly covers another what was going to be the consequences of of of say year city outlines stuff and that on layer of placing over on top of your rotor something like that where the courts where will matter where will it not and ending again the classification breaks how you use class of optical map server how use the class scheme in order to symbolize over different classifications and where you set your time scale ranges and and of course you know how you want to align weights and and and line styles you pick for emphasizing and and same for symbols know was the sizing other symbols which symbol and what the effects of shadows and so forth to help you emphasized and and then the optical note about employers years the start of the list of the map server of words that are affected by the making these classifications and the yellow ones are the ones I think for the most part the interest of primary reporters and the Greens are actually got that's max features of the layer level and so forth I think or more like reasons you don't need to know some of you can actually use shortcuts and like for example the max features decision I'm only going to show some of my going to show X number of features in this in this but at this time this layer and so forth depict note toward color I think there's a whole world of discovering more more about that here in this conference about I because she caught color perception and what the effects of our of color settings how they affect will how they how they presented the emphasis on with the colors and this 1 resource i've been investigating called I think it's
fairly well known and it's a worksheet uh you can perform summer for is doing in the them the I the sum over I was doing when the quantum of page before we can pick the number of of breaks and on what basis and in what color scheme on some
just a little note about labels is that that is this thing is that you like if you double you're points size fectly using 4 times as much of pixel space to represent the label and don't of so basically the I was studying the how to like those same labels in particular I am I am those are 1 of my the main goal of getting paid for this because the because the US city-names to me anyway I'm fairly familiar with them the absence wrong them i'll see it pretty fast and farm
like it is probably pretty artists in here this is how a simple scheme or I've collected the standard deviation of the populations of the US cities and um and used actually use the used as a standard deviation of a square root of the population value turned out to be a pretty good foundation for of differentiating the populations and the importance of the cities on the map and the another case I've had before where where doing a one-size-fits-all type classification scheme does it always work this is the city of Austin there where I live where I used a lot of the US 1400 class in the target dataset is basically residential and small you moderate use were all roads and so forth but they don't classify very well within that namely the what I call major thoroughfares are not separately classified you have to be either a pull them out through their own names such as you merely adds not reliable you can't say all the avenues of boulevards pathways are important roads so this I simply tried to have a length threshold is set any road long any as 14 and a road longer than I think 3 kilometers I will classify as a major thoroughfare happens to work in much of the United States particularly the South I anything until anything that's part outside the central Plains where the roads curve you get of minor roads there especially long but I just did a Portland map here you see the dark this contrast is pretty poor here but not all of the light rail what was what because her you might work well on on the island were successfully whereas over here in the older part of of town there's to me dark lines of many of ordinary thoroughfares or an or ordinary residential streets have to be as long as the thoroughfares in my scheme is breaking down here and so the point at point being is that it's in the cases that we try to do an analytical study about how to classify of of roads according to their length it's gonna vary from place to place and that a single scheme as often as not to work well across a border and I what I've I've been there of seen opportunities for her help to the use the analysis of the data I'm not only fora within a layer about how to classify between layers get some clues about about around about how you might wanna speculators and and and I I would also say opportunities for using so analysis to prescribe this scale breaks and and also time the that and so what I might have a point where I I rather than try to doctor the same map server internally to on to take on some of these kinds capabilities I start I think is more appropriate to try to model outside for example of what I'm learning more more about scribe here at this conference and it seems like is more suitable vehicle to package the statistical calculations about the data as part of its the process process and so if I were
to create a pie-in-the-sky description of a truly outside at that would help but I'm prescribed styles from of data analysis I would of course make it integrate well with the individual products like map server and and on and actually be able to I have access to the data source like if you're impose giants and you can actually if you need to for performance 6 a classification value of as that by adding a column to tape wouldn't have that you have the means right there to do so and and click into the the state of the art color perception and worksheets such as the 1 shown earlier the colorbar and and integrate well with the symbol creations of worksheet and Pat alignment styles of font worksheets and and can build a spy graphs of the of the of the of the queries that you're doing on your data and so on so so I
think you're much FIL the with and I just had a
common so some of that's on your wish list that is actually packets in the Mac manager and the Windows product or map server so is prepackaged so yeah thank you it
Umwandlungsenthalpie
Zentrische Streckung
Bruchrechnung
App <Programm>
Befehl <Informatik>
Statistik
Subtraktion
Gewicht <Mathematik>
Quader
Extrempunkt
Winkel
Klasse <Mathematik>
Kombinatorische Gruppentheorie
Fokalpunkt
Computeranimation
Hypermedia
Textur-Mapping
Spannweite <Stochastik>
Menge
Server
Kontrollstruktur
Wort <Informatik>
Softwareentwickler
Fehlermeldung
Task
Mapping <Computergraphik>
Konditionszahl
Widerspruchsfreiheit
Computeranimation
Instantiierung
Klasse <Mathematik>
Gewichtete Summe
Punkt
Extrempunkt
Gruppenkeim
Gleichungssystem
Symboltabelle
Extrempunkt
Technische Optik
Raum-Zeit
Computeranimation
Übergang
Homepage
Eins
Freeware
Font
Arithmetischer Ausdruck
Font
Kontrollstruktur
Gerade
Taupunkt
Cliquenweite
Schnelltaste
Zentrische Streckung
Statistik
Sichtenkonzept
Reihe
Nummerung
Kontextbezogenes System
Arithmetischer Ausdruck
Entscheidungstheorie
Spannweite <Stochastik>
Softwarewartung
Mustersprache
Randwert
Gruppenkeim
Menge
Konditionszahl
Server
Azyklischer gerichteter Graph
Dateiformat
Projektive Ebene
Tabelle <Informatik>
Zeichenkette
Nebenbedingung
Darstellung <Mathematik>
Total <Mathematik>
Gewicht <Mathematik>
Azyklischer gerichteter Graph
Algebraisches Modell
Klasse <Mathematik>
Zahlenbereich
Textur-Mapping
Pufferspeicher
Informationsmodellierung
Datensatz
Bildschirmmaske
Spannweite <Stochastik>
Hasard <Digitaltechnik>
Datentyp
Quantisierung <Physik>
Abschattung
Inverser Limes
FAQ
Gleichungssystem
Bildgebendes Verfahren
Hilfesystem
Soundverarbeitung
Trennungsaxiom
Green-Funktion
Hasard <Digitaltechnik>
Winkel
Mailing-Liste
Symboltabelle
Objektklasse
Partitionsfunktion
Mapping <Computergraphik>
Generizität
Mereologie
Graphfärbung
Wort <Informatik>
Kantenfärbung
Verkehrsinformation
Prozess <Physik>
Punkt
Klasse <Mathematik>
Raum-Zeit
Computeranimation
Textur-Mapping
Webforum
Parkettierung
Datentyp
Kontrollstruktur
Wurzel <Mathematik>
Kontrast <Statistik>
Kurvenanpassung
Gerade
Hilfesystem
Analysis
Beobachtungsstudie
Zentrische Streckung
Dicke
Schwellwertverfahren
Pixel
Nummerung
Rechnen
Sturmsche Kette
Mereologie
Server
Analytische Menge
WebDAV
Innerer Punkt
Standardabweichung
Azyklischer gerichteter Graph
Datenanalyse
Symboltabelle
Gerichteter Graph
Computeranimation
Datensichtgerät
Deskriptive Statistik
Textur-Mapping
Mailing-Liste
Font
Freeware
Font
Magnetbandlaufwerk
Datenmodell
Abfrage
Symboltabelle
Biprodukt
Gerade
Inverser Limes
Arithmetisches Mittel
Mustersprache
Server
Azyklischer gerichteter Graph
Kantenfärbung
Modelltheorie
Aggregatzustand
Textur-Mapping
Datenmanagement
Bildschirmfenster
Server
Vorlesung/Konferenz
Mailing-Liste
Biprodukt

Metadaten

Formale Metadaten

Titel Creating Map Style & Visibility Rules from Statistics
Serientitel FOSS4G 2014 Portland
Autor Hollingsworth, Robert
Lizenz CC-Namensnennung 3.0 Deutschland:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen.
DOI 10.5446/31978
Herausgeber FOSS4G, Open Source Geospatial Foundation (OSGeo)
Erscheinungsjahr 2014
Sprache Englisch
Produktionsjahr 2014
Produktionsort Portland, Oregon, United States of America

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Map style, label, and visibility rules, especially those aimed at differentiating "important" classes of features from "minor" ones, can be derived from statistical functions performed on feature attributes. If the source data classification scheme is not already strong in prioritizing features how we want to view them, then style patterns may emerge from calculations over an assortment of counts, sums, averages, and other measurements. We will begin with a quick examination of popular open source web and desktop mapping engines -- do their configuration capabilities include formal constructs for deriving rules from statistics? Or must the developer arrive at "this looks right" through trial and error? We'll extend the discussion to specific data distribution patterns that can be exploited for styling. We're accustomed to setting line styles, symbol and font sizes, colors, and visibility at different scales. The bell curve resulting from a query may point us to where we make the scale breaks, or toward how much color or size contrast to employ in order to make the best presentation from the particular data we are displaying. Perhaps we can arrange our queries, thereby grouping our features a certain way, to aim for an "ideal" curve that is already known to produce pleasing results.A simple set of query tools for streamlining style assists from statistics will be used to create a few examples from troublesome data.
Schlagwörter statistics
distribution curve
mapping
styles
configuration
visibility
classification
query tools

Ähnliche Filme

Loading...