Knowing the granularity of the data is crucial.įor more information, see the Free Training Video on aggregation and granularity (Link opens in a new window) or the Help topic Data Aggregation in Tableau. What does a row (aka record) in the data set represent? A person with malaria? A provinces' total cases of malaria for the month? That's the granularity. Granularity refers to how detailed the data is. You can change the aggregation to things like average, median, count distinct, minimum, etc. By default, measures in Tableau are aggregated.Aggregation and granularity are opposite ends of a spectrum.Īggregation refers to how the data is combined together, such as summing all the searches for Pumpkin Spice or taking the average of all the temperature readings around Seattle on a given day. Understanding aggregation and granularity is a critical concept for many reasons it impacts things like finding useful data sets, building the visualization you want, combining data correctly, and using LOD expressions. For example, you would be unlikely to find a data set with case-by-case reporting of malaria by address, so monthly totals by region might be granular enough. Note that due to privacy or practicality, some data sets will never be fully granular. What counts as disaggregated can vary by analysis. Ideally you want to get your hands on daily data, so you could see the huge spike when Starbucks starts offering #PSL. For example, if you want to look at trends in people Googling "Pumpkin Spice" but have yearly data, you can only look at a very high level overview. If the data is too aggregated, there isn't much you can do for analysis. A good data set is disaggregated (raw) data Some features or viz types may require specific characteristics of the data such as:Ģ. "Interesting" measures-either substantial variation in magnitude or positive and negative values.Not all data sets need all these elements-know what you need for your purpose and don't waste time with data sets that are missing key elements. Basic demos often involve drilling down into dates, so the data would need at least one date field (and it would need to be more granular than just year to show drill down). For example, maps are a great visual but require geographic data. If you're looking for a data set to build a specific visualization or to showcase specific functionalities, make sure the data set has the types of fields you need. A good data set has the elements you need for your purposes It's also clean data, with only the correct values in each field and a good data structure.ġ. Small and clean: Superstore is only a few megabytes so it takes up little room in the Tableau installer.You don’t need to look up what any values mean. Metadata: Superstore has well-named fields and values.There are also multiple measures and dates, which open the possibilities for chart types and calculations. Dimensions and measures: Superstore has several dimensions allowing us to "slice and dice" by things like category or city.Those items can be rolled up to the order level (via Order ID) or by any of the dimensions (such as date, customer, region, etc.) Disaggregated: The row-level data is each item in a transaction.There are very few chart types you can't make with Superstore alone, and few features it can't be used to demonstrate. Necessary elements: Superstore has dates, geographic data, fields with a hierarchy relationship (Category, Sub-Category, Product), measures that are positive and negative (Profit), etc. ![]() Superstore is one of the sample data sources that come with Tableau Desktop. Are useable (not in a proprietary format, too messy, or too cumbersome).Have good metadata or a data dictionary. ![]() Have at least a couple dimensions and a couple measures.On the whole, look for data sets that meet the following conditions: However, there are some considerations that can help you weed out data sets that are unlikely to suit your purpose. ![]() As long as that need is met, it's a good data set.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |