Data Application: Extract for mining, fusion and reporting of viewing behavior by demographic.
AMRLD is a files available from Nielsen holding respondent information and activities and derived from their Elemental Data Warehouse:
- Receive Nielsen AMRLD (All Minute Respondent Level Data) files.
- Process and load data to a RDBMS (Relational Database Management System).
- Process the data to retrieve common metrics (i.e. GRP, Reach, Impressions, VPH, etc.) by standard demographics.
Data Application: Gather data to support the analyses on task.
There are many different sources of data available for reporting. These data sources come in many different formats (i.e. RDBMS, Excel, Flat Files, etc.). Once the analyses are defined, we will gather the information necessary to support these analyses. This data will be translated into a format that allows for fast mining and analysis. This processing will be and should be transparent to our clients
Data Application: analyzing data from different perspectives and summarizing it into useful information
We can analyze data from many different dimensions or angles, categorize it, and summarize the relationships identified. Technically, data mining is the process of finding correlations or patterns among dozens of fields in large relational databases.
Data Fusion and List Matching
Data Application: Prepare disparate datasets for data mining.
This is the process of synthesizing raw data from several sources to generate more meaningful information that can be of greater value than single source data. Data fusion, list matching, is a subset of data integration. They are very similar in that they have the same end result. They are all used to couple data together. The way in how the data is joined and the source of the data is the differentiating factor. Data fusion uses disparate data sources and via statistical models 2 or more disparate lists are joined together to transpose behaviors across sources. List matching joins lists together that already have a known link between them, so no modeling is necessary.
Data Application: Collect, organize, interpret, and report data using a practical application of statistics.
o Raw data
Machine Learning / Statistical Modeling
Data Application: Sift through data to identify patterns which in turn will use to find other patterns, outliers, groups, etc.
o ANCOVA – Analysis of Co-Variance
o ANOVA – Analysis of Variance
o Neural Nets – A set of algorithms that emulate human behavior to predict future behavior.
o Regression Models – Looking at the relationship of a dependent variable to one or more independent variables. This is one of the main tools of predictive analytics.
o Clustering – Grouping of datasets to show correlation between the data points.
Customer Relationship Management (CRM)
Data Application: Utilize the customer database for additional KPI’s
Other Third Party Data
Data Application: incorporation of other data like: FanPage, FaceBook, Google, etc.
o Bring data into the mix allowing a look at data from different perspectives.
o Mixed Media reporting