#mapshaper

Ok. 

I wanna know why have I never heard of this online tool before. Like, what the hell is wrong with the social media? Is something wrong with Twitter or Instagram or something that they never caught on mapshaper? Or was it just me and my hazardous ignorance, yet again?

Have you tried this free nifty online tool that literally simplify crazy complicated shapefile polygons like it’s no one’s business?!

It started with some last minute inspiration on how to collate data from 3 different regions; developed from remote sensing techniques which vary from one another. The common output here is to turn all of them into a vector file; namely shapefile, and start working on the attribute to ease merging of the different shapefile layers.

Once merged, this shapefile is to be published as a hosted feature layer into the ArcGIS Online platform and incorporated into a webmap that serves as a reference data to configure/design a dashboard. What is a dashboard? It's basically an app template in ArcGIS Online that summarizes all the important information in your spatial data. It's a fun app to create, no coding skills required. Check out the gallery here for reference:

Operations Dashboard for ArcGIS Gallery

There are two common ways to publish hosted feature layer into ArcGIS Online platform.

Method 1: Zip up the shapefile and upload it as your content. This will trigger the command inquiring if you would like to publish it as a hosted feature layer. You click 'Yes' and give it a name and et voila! You have successfully publish a hosted feature layer.

Method 2: From an ArcGIS Desktop or ArcGIS Pro, you publish them as feature service (as ArcMap calls them) or web layer (as the its sister ArcGIS Pro calls them). Fill up the details and enabling the function then hit 'Publish' and it will be in the platform should there be no error or conflicting issues.

So, what was the deal with me and mapshaper? 

🛑 A fair warning here and please read these bullet points very carefully:

I need you to remember...I absolve any responsibility of what happens to your data should you misinterpreted the steps I shared. 

Please always  👏🏻  BACK 👏🏻 UP  👏🏻  YOUR 👏🏻 DATA. Don’t even try attempting any tools or procedure that I am sharing without doing so. Please. Cause I am an analyst too and hearing someone else forget to save their data or create a backup is enough to make me die a little inside. 

For this tool, please export out the attribute table of your shapefile because this tool will CHANGE YOUR SHAPEFILE ATTRIBUTES. 

When I was publishing the vector I have cleaned and feature-engineered via ArcGIS Pro...it took so long that I was literally dying inside. I'm not talking about 20 minutes or an hour. It took more than 12 hours and it did not conjure the 'Successfully published' notification as I would've expected from it.

So at around 5.30 am, I randomly type 'simplify shapefily online free'. Lo and behold, there was mapshaper.

All I did was, zip up my polygon, drag it to the homepage and it will bring you to the option of choosing the actions that will be executed while the data is being imported into mapshaper:

detect line intersections

snap vertices

This option will help you to detect the intersections of lines within your vector/shapefile. This can help identify topological error.

The option to snap vertices will snap together points of similar or almost identical coordinate system. But it does not work with TopoJSON formats.

There is something interesting about this options too; you can enter other types of customized options provided by the tool from its command line interface! But hold your horses peeps. I did not explore that because here, we want to fix an issue and we'll focus on that first. I checked both options and import them in.

This will bring the to a page where there you can start configuring options and method to simplify your vector.

To simplify your shapefile, you can have both options to prevent the shape of the polygon being compromised; prevent shape removal, and to utilize the planar Cartesian geometry instead of the usual geoid longitude and latitude; use planar geometry. The implication of the second option is not obvious to me yet since all I wanted was to get the data simplified for easy upload and clean topology, thus, I chose both options to maintain the shape and visibility of all my features despite the highest degree of simplification.

Alike to the options of methodology for simplication in the mainstream software, I can see familiar names:

Douglas-Peuker

Visvalingam / effective area

Visvalingam / weighted area

First and foremost, I had no slightest idea of what these were. Like for real. I used to just go first for the default to understand what sort of output it will bring me. But here, the default; Visvalingam / weighted area, seemed like the best option. What are these methodologies of simplification? There are just algorithms used to help simplify your vectors:

🎯 Douglas-Peucker algorithm decimates a curve composed of line segments to a similar curve with fewer points (Ramer-Douglas-Peucker algorithm, Wikipedia; 2021).

🎯 Visvalingam algorithm is a line simplication operator that works eliminating any less significant points of the line based on effective area concept. That basically means that the triangle formed by each of the line points with two of its immediate neighboring points (Visvalingam Algorithm | aplitop).

🎯 Visvalingam algorithm with weight area is another version of Visvalingam algorithm of subsequent development where an alternative metrics is used and weighted to take into account the shape (Visvalingam & Whelan, 2016).

For reasons I can't even explain, I configured my methodology to utilize the third option and now that I have the time to google it, Thank God I did.

Then, see and play with the magic at the 'Settings' slider where you can adjust and view the simplification made onto the vector! I adjusted it to 5%. The shape retained beautifully. And please bear in mind, this vector was converted from a raster. So, what I really wanted is the simplified version of the cleaned data and to have them uploaded.

Now that you've simplified it, export it into a zipped folder of shapefile and you can use it like any other shapefile after you extracted it.

Remember when I say you have got to export your table of attributes out before you use this tool? Yea...that's the thing. The attribute table will shock you cause it'll be empty. Literally. With only the OBJECTID left. Now, with that attribute table you've backed up, use the 'Join Table' tool in ArcGIS Pro or ArcMap and join the attribute table in without any issues.

Phewh!!

I know that it has alot more functions than this but hey, I'm just getting started. Have you ever done anything more rocket science than I did like 2 days ago, please share it with the rest of us. Cause I gotta say, this thing is cray!! Love it so much.

mapshaper developer, if you're seeing this, I 🤟🏻 you!

UPDATE

I have been asked about the confidentiality of the data. I think this is where you understand the reason behind the fact that they will work even with using just the ‘.shp’ file of the shapefile since _that_ is the vector portion of the shapefile. 

Shapefile is a spatial data format that is actually made up of 4 files; minimum. Each of these files share the same name with different extensions; .prj, .shx, .shp and .dbf. Although I am not familiar with what .shx actually accounts for, the rest of them are pretty straightforward:

.prj: stores the projection information

.dbf: stores the tabulated attributes of each features in the vector file

.shp: stores the shape/vector information of the shapefile.