Browsing by Author "Kupsu, Mikko"
Now showing 1 - 3 of 3
- Results Per Page
- Sort Options
- Evaluation of Big Data SQL frameworks for ad-hoc analysis of digital audience measurement data
Perustieteiden korkeakoulu | Master's thesis(2017-05-08) Sundarraman, SridharBig Data analytics has now become quintessential for information exploitation, given the amount and rate at which data is generated everyday in the world. One of the key fields in which Big Data analytics has immense benefits is digital audience measurement. Ad-hoc analysis of digital audience measurement data helps model specific user behavior and provides precise insights for product development and brand engagements. This thesis focuses on an evaluation of contemporary Big Data processing frameworks that support SQL, and assesses their applicability and effectiveness for such ad-hoc analysis purposes. The chosen representative Big Data SQL frameworks, namely Apache Hive, Apache SparkSQL, Facebook Presto and Amazon Athena, are mainly evaluated in terms of their performance and cost effectiveness. To this effect, we have devised 12 different workload queries that are executed on varying sizes of raw HTTP measurement data stored as Parquet files in AWS S3. For each combination of dataset size (ranging from 70 GB to 1.7 TB) and workload query, response times of the four frameworks are measured. These tests are performed on an AWS Elastic MapReduce (EMR) cluster for Hive, SparkSQL and Presto, while Athena being a managed Big Data SQL service does not require any infrastructure to be setup. In addition to the performance and cost effectiveness, we also present a comparison of these frameworks in terms of their usability and language flexibility they offer. The results of this evaluation shed light on the strengths and weaknesses of each of the framework in various aspects of comparison and their overall suitability for ad-hoc analysis on raw audience measurement data. - Map overlay - analyysimenetelmä ja sen sovellusmahdollisuuksia
Insinööritieteiden ja arkkitehtuurin tiedekunta | Bachelor's thesis(2009) Kupsu, Mikko - Ajantasaisen maastotietoaineiston tuottaminen karttatuotannon ja tietotuotepalveluiden käyttöön
School of Engineering | Master's thesis(2013) Kupsu, Mikko