ArcCollector 2

INTRODUCTION

This lab explores geodatabase creation and data collection using ArcCollector. Unlike the previous ArcCollector lab, the area of study and theme of the project is decided by the student. The geospatial question explored in this lab was: what areas of the Third Ward in Eau Claire affected most by potholes?

An overview for the flow for the answering this question is as follows: A geodatabase was created with suitable domains, data was collected, data was brought into ArcMap, data was interpreted.

Data was collected on the following variables for each pothole:

• Location
• Size (in cm)
• Cluster size
• Overall road condition
• Area type (Residential, commercial, etc.)
• Traffic amount

While the data collection and interpretation went smoothly, there were some concerns that arose during interpretation. This will be discussed later.


STUDY AREA

The area of interest is a section of the Third Ward in Eau Claire, Wisconsin. This area was chosen because of the diverse road conditions, the mixture of commercial and residential properties, and the density of potholes. The study area is shown below in figure 1. The Third Ward is considered the downtown of Eau Claire. It is largely small businesses such as restaurants, bars, banks, etc. On the outer edges, particularly on the southern side of the area studied, is more residential. There is a mix of high and low traffic streets.

Figure 1: Study area

METHODS

An overview of the workflow was: geodatabase creation, setting up domains, creating fields and assigning domains, data collection, and data interpretation.

The first step was geodatabase creation. A suitable folder for the project was made, and a geodatabase was created using ArcCatalog. Next, still using ArcCatalog, domains were created. A screenshot of this menu appears below in figure 2. For each data type needed, a domain was necessary to determine the allowable values for each variable. This would minimize error. For example, the traffic amount was assigned as text, and options were given so no text had to be manually entered. Similarly, for number variables, such as pothole size, a range of allowable values were given.

Figure 2: Creation of domains within the geodatabase

Once the domains were created, a feature class was created for data collections. This feature class was assigned as points, which would be the locations of potholes. WGS1984 Auxiliary Sphere was used as the coordinate system, as this is necessary for data collection in ArcCollector. Fields were added to this feature class, and the previously created domains were assigned to fields. After this was completed, the feature class was uploaded to ArcGIS Online so it could be accessed in ArcCollector.

With the ArcCollector app installed on the iPhone used for this lab, the feature class was accessed by logging into an ESRI account. With the collection device ready, all streets within the study area were navigated and potholes were documented. An hour was chosen that would have little traffic, so the location could be collected more accurately by standing directly next to the pothole, as shown below in figure 3. For potholes on busier streets, however, the location was collected from the sidewalk. A few meters difference would have a negligible effect on the results, as it is generally within the uncertainty of the GPS signal anyway.

Figure 3: Data collection

Two screenshots appear below in figure 4 and 5 to show the interface of the ArcCollector app. When a pothole was found, the plus icon at the top of the screen was used to create a new data point. This would bring up the screen shown in figure 5, where variables could be assigned for the new data point. For variables such as TrafficRank, a list of allowable variables was shown. For number variables such as size, the value had to be within a certain range. If the value was outside this range, an error would be given.

 Figure 4: Map view within ArcCollector             Figure 5: Collecting data for a single point

Once all the data was collected, ArcGIS Online was used to download the data to ArcMap Desktop.



RESULTS

Shown below in figure 6 is a density map of potholes. The areas affected most by high pothole density are the streets with the highest traffic. This is likely due to a higher frequency of cars driving over the street causing an increased wear on the road surface.

Figure 6: Pothole density

Shown next in figure 7 is the distribution of average pothole sizes. Notice the largest potholes tend to be on the northeastern side of the study area. This area is residential with very little traffic. This makes sense, as less frequented streets will have a lower demand for fixing problematic potholes, so potholes in these areas could grow larger before there is a large demand to fix them.

Figure 7: Average pothole sizes

CONCLUSION

This lab proved to be an excellent learning opportunity. Much was learned about the theme and subject matter that can be collected and interpreted with the assistance of ArcCollector. This lab showed that data with discrete locations and mainly categorical variables are not ideal for interpolation. Turning discrete locations into continuous surfaces demands a range of numeric variables, as was learned in this lab. Adapting traffic data based on three categories (low, medium, high) into a continuous surface map is not ideal. If this lab were to be redone, a different theme would be chosen to study a variable with a more continuous nature.

Introduction to ArcCollector

INTRODUCTION

This lab explored ArcCollector, an ESRI app used to collect field data. ArcCollector can use crowd-sourced data and runs on any smartphone. Using a basemap, field measurements can be recorded with locational information taken through the phone's GPS. This can be used to create feature classes in the field. This data is automatically uploaded to ArcGIS Online, allowing other users to access it in real time.

In this lab, a series of attributes were created in advance, including temperature, dew point, wind speed, windchill, and wind direction. These measurements were recorded at many points throughout the study area. Using these data points, microclimates were mapped and assessed.


STUDY AREA

The study area of this lab was UWEC campus in Eau Claire. Prior to collecting the data, the study area was divided into 7 zones. The class was split into groups, each assigned to a different zone for data collection. These zones are shown below in figure 1. The zone I was assigned to was zone 3. This included the bridge connecting zones 1 and 2. Note that the Western section of zone 3, from the bridge over, was under construction and access was not possible. Measurements were taken on an overcast, ~50°F day.

Figure 1: The study area, divided into 7 zones used in data collection

METHODS

The equipment used in this lab were an iPhone, a Kestrel 3000 handheld weather station, and a compass for determining wind speed. Two people were assigned to each zone, so I had one other person collecting data in zone 3.

To collect data, the ArcCollector app was used with the project opened. For each data point, measurements were read off the Kestrel and entered into a new data point on ArcCollector. Temperature, dew point, and windchill were entered directly. To find wind speed, the Kestrel was held vertically, perpendicular to the wind direction. Once the wind speed stopped climbing, the measurement was taken. The direction of the wind was measured using a compass. The Kestrel is shown below in figure 2.

Figure 2: The Kestrel handheld weather station measuring wind speed

In figure 2, wind speed is being measured through the use of the fan. Notice the coil underneath the fan. This collects temperature information. Cycling through the available measurements was done using the arrow buttons.

Measurements for each attribute were held to domains assigned to the feature class before collection began. Examples of these domains are limiting wind direction from 0 - 360 degrees, limiting temperatures to only reasonable values, etc. This was done to prevent some user error, for example if someone hit an extra number when entering temperature, entering it at 500°F instead of 50°F, an error window would come up to tell them there is incorrect data entered.

As data was collected, it was uploaded in real time to the database. This was easily seen in ArcCollector, with new data points popping up across the study area from each classmate collecting new points. Once data collection was complete, the data was downloaded onto ArcMap and surface maps were created.


RESULTS

ArcGIS online was used to download the data to the desktop ArcMap. The collected data contained 239 points. Interpolation was needed to create a continuous surface to better represent the data. Several methods were tried, and the kriging method turned out to be the best option. The results from kriging showed more detail than other methods tried. The first map created was a temperature map, shown below in figure 3. The maximum variance in temperature was less than 7°F. As mentioned, the day was overcast, so there were no drastic variances from sunlight and shadows. A sunny day would likely create more interesting data. That being said, there are some slightly warmer sections of the campus shown below, mainly in pockets along the southern edge of the study area. Note that the study area shape has been simplified. This was done to make the surface gradient seem less choppy and disconnected.

Figure 3: Temperature map of UWEC

The next map created, shown below in figure 4, is the windchill map. Notice the similarity between this and the temperature map. They are almost identical besides the slightly lower minimum temperature on the windchill map. Warm and cold pockets are still found in the same areas with similar shapes and magnitudes.

Figure 4: Windchill map of UWEC

The next map created was the dew point map, shown below in figure 5. This map has similar distribution to the previous two, but is not identical. Notice two warmer pockets are found on the eastern section, similar to the other maps. The pocket on the southwestern section does not appear in the dew point map, however.

Figure 5: Dew point map of UWEC

The last map created was the wind map, shown below in figure 6. To create this map, raster surface maps were created for both wind speed and direction. This was done using the same kriging method as the previous maps. Once these were created, The raster was symbolized using symbols instead of a gradient. Once the direction was symbolized as arrows, the wind speed raster was used as a reference for the arrow magnitude. Notice the wind direction is mostly headed northwest, with the exception of some areas near the southern section. This could be explained by the hill that is found there. As for wind speed, the largest magnitudes are found in the western section. This is likely because this part of campus is more open, with less buildings and trees to block wind.

Figure 6: Wind speed and direction

CONCLUSION

This lab served as an introduction to ArcCollector. This app will be explored in more detail in a subsequent lab. ArcCollector proved to be a powerful, yet simple tool for data collection. Using smartphones as a data collection device is a creative solution, as smartphones are readily available and have powerful GPS capabilities. It should be mentioned how useful it was for domains to be set up before data collection. This definitely cut down on user error and simplified the process. The app itself was easy to use and allowed the data collection from other groups to be seen in real time.