CODEWITHSUNDEEP
beast
Ga1acy
Ga1acy

Reputation: 1

How to convert RDD[(String, int)] to RDD[IFeature] in beast? #beast #ucrbeast

I want to generate an image to show statistics on crime situations in different areas of a city,I chose to use zip code to represent different areas and use PBSM spatial join with Contains Predicate.

However I ran into some other problems, I tried to use countByKey to create zip code-crime pair, then I found that this is not an RDD, thus can't produce the spatial file, so I used reduceByKey, then created an RDD[(String, int)]. Now, I'm struggling with how to convert RDD[(String, int)] to RDD[IFeature] so that I can produce a shapefile and visualize it. I tried to use Feature.append() and Feature.create(), but I can't use it well without some documentation. I'm not sure if I'm in the right direction.

 try {
      // Import Beast features
      import edu.ucr.cs.bdlab.beast._
      // TODO Insert your code here
      val zipCodeBound = sparkContext.shapefile("/Users/galacy/beast-start-up/beast-project/src/Boundaries - ZIP Codes.zip")
      val crimePoint = sparkContext.readCSVPoint("/Users/galacy/beast-start-up/beast-project/src/Crimes_-_2001_to_Present.csv", "Longitude", "Latitude",',', skipHeader = true)


      val sjResult  = zipCodeBound.spatialJoin(crimePoint, ESJPredicate.Contains, ESJDistributedAlgorithm.PBSM)

      val mapRes = sjResult.map(f => (f._1.getAs[String]("zip"), 1)).reduceByKey(_ + _)
      val finalRes = mapRes.map(f => (Feature.create()))

    }

Upvotes: 0

Views: 11

Answers (0)

Related Questions