There are several parts of your setup that aren't entirely evident from your query. Tableau extracts are a handy tool since they are effectively a temporary cache of query results that persists. In many ways, they behave like a materialised view.
You'll almost always want to keep your extract in a central location, such as Tableau Server, so that it can be shared by a large number of people. That's not unusual. With a little effort, you can produce a duplicate of the extract for each Tableau Desktop user (say by distributing packaged workbooks). This makes sense in some situations, such as with distant users, but it is not the norm. This use case is comparable to monthly data marts being sent out to analysts using data from a central warehouse.
So, in response to your query, Tableau provides tools that you may utilise as needed to best suit your specific use case — either duplicated or shared extracts. The idea is to understand how extracts function and then use them as needed.
The simplest approach to distribute an extract is to publish it to Tableau Server, either as part of a workbook or as a standalone data source (which is then referenced by workbooks). After you've made an extract, the quickest method to repeat it is to export your workbook as a packed workbook.
A Tableau data source is meta data that refers to an original source, such as a CSV file or a database. An extract that shadows the original source can be included in a Tableau data source. To see new data, you can refresh or add the extract. You may have the refreshes happen on a schedule if you publish to Tableau Server.
It's better to save the extract centrally on Tableau Server, especially if the data changes infrequently. You may save query results, offload work from the database, save network traffic, and speed up your visualisations by capturing query results.
Filtering (and even aggregating) extracts to have only the data needed to present your viz can boost speed even further. Aggregation may be done once at the time of extract production, which is very handy for huge data sources like web server logs. Instead of repeating long-running SQL queries at visualisation time, extracts may just record the results.
If you use aggregated extracts, be sure that any further aggregation in the visualisation makes sense. MINS of MINS and SUMS of SUMS are well-defined terms. Averages of Averages, for example, are not always useful.