Thunderhead Explorer

Data Scientists, if you are processing and analyzing spatial data and are using Python , then ArcPy should be included in your arsenal of tools and ArcMap should be utilized for geo spatial data visualization.  Following the last post where I extended Apache Hive with spatial User Defined Functions (UDFs), in this post I will demonstrate the usage of the "extended" Hive within Python and how to save the output into a feature class for rendering within ArcMap or any ArcWeb client using ArcGIS Server. Given a running Hadoop instance and assuming that you have installed Hive and have created a Hive table as described in the last post, start the Hive Thrift server as follows: $ hive --service hiveserver When ArcGIS for Desktop is installed on a host, Python is optionally installed and is enabled with GeoProcessing capabilities. Install Hive on your desktop and set the environment variable HIVE_HOME to the location where Hive is residing. To access the Hive python ...

Following the last post , where we used Pig  to analyze data stored in HDFS, in this post we will be using Hive and spatially enabling it for geo analysis. Hive enable you to write SQL like statements in a language called HiveQL that Hive converts to a MapReduce job that is submitted to Hadoop for execution. Again, if you know SQL, then learning HiveQL is very easy and intuitive.  Hive is not intended for OLTP and real-time analysis. Like Pig, Hive is extensible via User Defined Functions (UDFs), so we will be using almost the same functions as in the previous post to find things that are near and/or within some criteria. There are several ways to store data in HDFS, one of them is in the SequenceFile format. This is a key/value binary format with compression capabilities. For this post, I will be transforming an input into a well know binary format to be stored onto HDFS for later query and analysis. An object that required persistence onto a SequenceFile has to i...