User Tools

Site Tools


Meta Data


Meta data is information about data. If you store a PDF-File the text in the File is data. The fact that this file is a PDF-Document is meta data as well as the creation date or the user who created the file for example is meta data. Meta data is often used to describe the data without manipulating the data. Meta data describes the data in general not in detail. In business cases meta data can also be used for legal purposes for example for example when which employees did missed a fatal error or debt.

Usage of meta data

In the past meta-data was only used to find data quickly or to tell the viewer how to read the data. So your google rank for example was nearly only calculated by the meta data you implemented to your website. SEO (search engine optimisation) was in the late 1990`s nearly only made via meta data. Today the algorithm of search engines is much better and the meta data lost relevance for search engines placement in the last few years.

Meta data in Business Intelligence

In Business Intelligence we try to get information as fast as possible in the best case our information will be delivered in real time. But what if we have a Big Data scenario where the Data comes from different sources for example from a sql database, from email and from social media. In this case we have a large pool of data witch we have to go through. Here meta data can us to prefilter the data we have to go through quickly. Meta data often is at the beginning of the data in the head of the data file/set so we can quickly sort our data by only reading the headers. After we filtered the data we have less data to analyse or to work with. The less data we need to analyse the more likely we are approaching the goal of real-time analysis.

On the other hand meta-data makes data readable and understandable for example the social media data if we get following dataset from a social media platform:

{“Alan”;”Dave”; “02-03-04”; “marketing”; “finance”}

If we try to analyse this we have several problems like what’s the name and what`s the first name, what means this date and is it the second of March or the third of February and finally what does marketing and finance mean?

If we get this set witch meta data like the following example the data will be readable. For more than one Dataset we can use the same meta data information.

{“Alan”;”Dave”; “02-03-04”; “marketing”; “finance”}
{“Mike”;”Muster”; “03-02-02”; “mechanic”; “computers”}

Meta data in the future

In the future there will be so many data source with different meta data that it would be nearly impossible to train your system how to read meta data. For this case it is necessary to standardize meta data in global.

Sources 13.04.2016 16:51 14.04.2016 08:30 14.04.2016 09:45

concepts/meta-data.txt · Last modified: 2020/08/20 13:27 (external edit)