Lab 6: ArcCollector Data Collection - Street Lights of UWEC off Campus Housing Area.

Introduction

In the previous lab, students used ArcCollector to collect microclimate data in small groups of 2-3 during a portion of a class period.  This activity was mainly tailored to get students familiar with the interface of the mobile software while also exemplifying how easy the process of combining the classes various data sets was.  Prior to this activity (lab 5) being started, a geodatabase was created for students which was then exported to ArcGIS online and accessed through collector - students had to do very little initial set up prior to collecting data.  In this lab, however, students will have to formulate their own geodatabases and collect their own data with the intent of illustrating a certain geographic trend or answer a particular question, with a geographic scope.  

The power of Arc Collector is that it is intertwined with ArcGIS online which allows users to connect with other users to share data.  Organizations then have the capability to collect data and instantly share it to corresponding parts of their operations, provided that they have access to the same ESRI group account.  As the account manager, one can create groups that field workers can then access and interact with maps/features that have been previously created.  

The specific question that is going to be answered in this lab via the use of ArcCollector is where are the street lights in the Randall Park neighborhood that are not working, faulty, obstructed, dammaged, or dimming relative to the majority that are fully functioning?  This question is important because having a well lit neighborhood can prevent crime from occurring, as crooks tend to avoid places where they can be identified or spotted.  Street lights failing is not an uncommon issue within the city of Eau Claire, and the Randal Park Area is of no exception. Sighting the City of Eau Claire web sight, this is the first paragraph on their page about street lighting:

"Reports of non functioning lights, commonly found in residential areas, which are mounted on a wooden pole, should be directed to Excel Energy by calling 1-800-628-2121 or by going to http://www.xcelenergy.com/Outages/Report_Outage "

What this implies is that the city is not well equipped with the resources to handle such issues, and thus has the cites major energy provider deal with this reoccurring issue.  In theory, this application could be a tool help identify faulty street lights. Going forward, This lab will walk through the steps of how this ArcCollector application was created, and will also present the data collected in an embedded map, followed by a discussion of the findings.

Study Area

The study area for this application will be the Randall park neighborhood located across the the Chippewa River from the University of Eau Claire main campus.  The area is primarily residential, populated with low value college properties, with a few pockets of nicer homes.  Below, in figure 1, one can see a map of this area.  Notice that the streets run very straight and form a uniform block system throughout the neighborhood. 

figure 1: Randall Park Neighborhood and the localized area of interest for street light data acquisitions  

Because the Randall Park Neighborhood is rather large for one person to collect street light data, the AOI was appropriately  localized to a smaller extent within the neighborhood.  As such, this localized section of the neighborhood is roughly 3 blocks long, beginning at Hudson St. and going south to Chippewa St, and roughly 7 blocks wide, beginning at 1st Ave and going west to 8th Ave. Let us now move on to how this application was produced. 

Methods

To discuss all the parts of this lab that brought this application to life, the methods section will be divided into a pre-publishing and post-publishing section.  The pre-publishing portion will discuss what went into creating the features before they were exported to ArcGIS online and used in ArcCollector. Post-publishing will discuss aspects related to the data acquisition process. 

Pre Publishing

Prior to this app being created, there is much to consider in terms of how to go about creating this application. The first step is to create the appropriate feature class and populate that feature class with the appropriate field information.  To get a better idea of what fields should bed created, it was necessary to look into what common damages occur to street lights and what issues tend to arise from those damages.  Referring to the link mentioned in the introduction of this lab (  http://www.xcelenergy.com/Outages/Report_Outage) the more common issues that occur with street lights tend to be as follows:

• light out 
• light on and off 
• light on during the day 
• light is dim 
• globe is broken
• globe is hanging 
• light pole is broken 
• light pole is leaning

Going off of this information, five fields were created in association with  this feature class that go off of the information provided by the Excel Energy damage web-form.  Four out of the five fields created make use of domains as limiting factors. The one that has no domain (obstruction_comments) was left as text field without domains so that users could be specific when describing the nature of obstructions to street lights.  Three other Fields were assigned domains, these can be seen below in figure 2 in the section titled default values and domains.

figure 2: domain and sub-type information 

Two types of domains were used for these fields that were apart of the street lights feature class - range and codded value domains. Here is a description of the domains used

Mnt_type: Describes the material that makes up the pole (wood or metal)

Glow_range: A range value that will be used to help identify dim lights, quantitatively (0-100 meters)

Globe: Provides categorical options that describe the globe outside of the bulb (intact, cracked, or hanging off)

Along with using domains, the final field that was created was made as a short integer so that it could be used as a sub-type, even though its value in the application is categorical and not numerical. The field light_type was chosen to be a subtype mainly to provide a variation in symbology that best represents the different types of lights that are used throughout the AOI.  The two options for this sub type are orange light and white light.  Orange light tends to be used in residential areas with low traffic and white lights tend to be used more in areas where there is a higher amount of car traffic.  Once the feature class was created and assigned appropriate symbology for each sub type, it was time to publish the map to ArcGIS online and begin taking point locations of street light locations.

The reason these fields were employed was in order to create a simple set of features that could encapsulate the status of a street light. The most important field is the quantifiable glow radius, as this allows for the identification of lights that are under performing either because of obstructions or defections within the fixture itself.  With appropriate symbology, this can directly show viewers the lights that are not adequately lighting their surrounding areas. Not only that, simply having the geographic locations shown also shows how their are certain block sections that have a single light, or sometimes none at all.

Post Publishing

Once published, the data was ready to be collected.  Data was collected was on the night of Sunday the 4th of April, 2016.  The data collection process took  roughly 2 hours, starting at 9 pm and ending shortly after 11 pm.  A bike was used as the the means of transportation throughout the neighborhood, from light to light. The night itself was rather damp and cold, hovering around 20 degrees, with high winds.  That being said, the default values, domains and sub-types allowed for the data collection process to happen relatively smoothly/quickly and as a result, the cold did not to need to be endured for too long.   More than 40 street light locations were recorded during this two our period.  The data was collected on Galaxy Prime - an Android product operating on a T - Mobile 3G network.

Results 

Below is an embedded map of the data that was collected on Sunday the 3rd of April, symbolized by the glow radius of the light on the ground and light type.  The map also has several toggle-able features, including a legend, zoom bar, address search bar, and base map selector.  Also, by clicking on the individual features, a pop window will appear showing the details about that specific street light 

If one clicks on the majority of street lights with smaller ground light radii, they have some sort of associated damage to the globe or obstruction that could account for their smaller light spread.  However, there are also several lights  that are simply dim with no sort of visible damage or obstruction from any outside objects.


As for the obstructions, the only obstructions that were observed in this study were caused by tree branches.  This perhaps does not have so much of an effect now, in early spring, but very well could have one in later seasons when these branches grow leaves and will have more material which could block light from shedding through.

For the most part, very little damages were found, yet that is not much of a surprise since there is a known forum which people can report damages directly too the company in charge of their maintenance.  That being said, the main damages that were found were cracked globes that surrounded the light, but nothing extensive to the point where a light was not functioning or flickering.  Below is  map that shows the lights that are damaged amidst the ones that are intact and functioning.  Like in the previous map, clicking on the feature will open up a pop up window that provides further detail about that individual street light

Discussion 

Working with ArcCollector going forward will be very useful tool as a field tool.  If using this, however, its very important that the individual taking data have a very solid understanding of the ins and outs of the subject matter.  If not, there is a chance that certain aspects of the collection process and pre-processing will be done incorrectly, redundantly, or be unnecessary. My knowledge of street lights is very little, and the tools that were at my disposal was essentially my phone and my person.  I thought that measuring the ground radius of light that shined onto the ground from the light fixture would be straight forward and an adequate way to to measure the functionality of the light.  However, I came to find out that the area that the light's glow was not uniform, as many of the poles are leaning in one way or another, causing the lights ground glow shape to be skewed or uniform in such a way that made it hard to measure definitively.

Although these results do not yield anything too note worthy in itself, imagine if this was a study that was done over a broader area, over a larger period of time, and/or with more than one field worker inputting features on the system.  This study was done in two hours with bike.  With a car, or two, with some more advance measuring equipment, these same results could be produced faster, with more meaningful results.  This is meant not just in the case of measuring the functionality of street lights, it could just as well be applied to measuring other infrastructure like telephone poles, street signs, road ways, bridges, and many other things.  The potential is very high. On top of the actual data collection.  ArcGIS online resources have the capability of being optimized further by employing custom widgets and tools via the use ArcGIS API java script.  For example, a query widget in this map would allow for viewers to look for certain traits contained within the fields, or features that occupy a certain area.

Conclusion 

Within this small area in the center of city of Eau Claire, the Randall Park Neighborhood, most of the lights are functioning just fine.  With the exception of three lights that had cracks and minor obstructions, most of the lights were found to functioning with little hindrance to the ground below.  It was not expected that this study would reveal that half of the neighborhood had lights that did not work, and the findings a of a just a several broken globes and obstructions form objects near the light was to be expected. The results, however, in this instance is not necessarily the focal point of this lab, even though the question was posed as to are there lights under preforming within the area of interest specified.  The broader goal was exemplify how seamless and efficient this technology can be when pre-processing is done with careful consideration.  A forum like what was presented in the link to Excel website could work in many locations, but in an area predominately occupied by younger people, students, and people of low income, its hard to imagine that many would use the sight.  As such, the absence of good lighting could go unreported and as a result, crime could become attracted to these areas with low viability. Utilizing this technology could be a proactive way to identify areas where lights are not functioning or under performing.