4. How to use Anatella?
4.1. A first example.
4.1.1. About the “Data-Preview” window
5 min
you can copy paste the results that are inside the “data preview table” into excel for example open excel and paste the results (press ctrl v) you obtain the header of the “data preview” table contains the number of rows inside the table at this output pin the number of columns inside the table at this output pin the status of the output table and the status of the view this status can be “ complete” the table at this output pin has been completely computed (it will thus been used to accelerate further computations along this data flow see section 4 2 6 for more information about this subject) and the current “data preview” window is able to directly show you the complete result table “complete (partial view)” the table at this output pin has been completely computed but only a partical view of this table is visible because the table is too big to show in its entirety “fragment” only a fragment (i e the first few rows) of the table at this output pin has been computed (typically because the user clicked the stop button on the toolbar to abort the computations before completion) this fragment is now visible inside the “data preview” note an incomplete table cannot be used to accelerate the computations later on you can “hover” with your mouse above a cell (or above a column title) to see the complete cell (or to see the complete title) for example you can quickly obtain a “data preview” of the data at any point inside your anatella data transformation graph (there is one exception see section 5 26 2 ) to know how to create and use “data previews”, see the section 4 2 6 the “data preview” window is thus a great tool to check for data quality issues check if you correctly parametrized your actions check if “everything goes as expected” in general the “data preview” allows you to visually inspect row by row all your tables but, sometimes, this is not enough you want to have a visual summary of the content of a column (just to be sure this column does not contain a small quantity of absurd values) you can easily get this summary by right clicking the header of a column for example, when i right click the header of the column “age”, i obtain when i select inside the above context menu the option “create chart as value var”, i obtain a new window from this point, i can either completely abort the chart computation click the button click the “stop” button to compute the charts on the rows that have already been processed the quantity of rows already processed is displayed here “199523” simply wait for the complete table to be processed to start a new “chart computation”, you don’t need to wait for the complete table to be computed (i e you don’t need to wait for the data preview status to reach “complete”) you can directly start the computations while the status is still on “fragment” this allows you to quickly produce some charts on the first few rows of your table no need to wait! the charts are produced as ms word and/or ms excel documents here is an example on the classical “census income” database let’s assume that we defined as “binary target” the column “taxable income amount” we asked to see the column “age” as a value variable then, we obtain univariate auc= 57 11 % count % count proba % dataset dataset target target index 6 3 % 6 3 %( v<= 3) 12492 6 3 0 0 0 0 0 12 8 % 12 8 %( 4<=v<= 7) 13039 6 5 0 0 0 0 0 19 1 % 19 1 %( 8<=v<= 11) 12611 6 3 0 0 0 0 0 25 2 % 25 2 %( 12<=v<= 15) 12206 6 1 0 0 0 0 0 32 9 % 32 9 %( 16<=v<= 21) 15323 7 7 28 0 2 2 9 39 8 % 39 8 %( 22<=v<= 26) 13642 6 8 220 1 6 26 0 45 8 % 45 8 %( 27<=v<= 30) 12024 6 0 566 4 7 75 9 52 5 % 52 5 %( 31<=v<= 34) 13368 6 7 1097 8 2 132 2 59 2 % 59 2 %( 35<=v<= 38) 13358 6 7 1534 11 5 185 0 65 4 % 65 4 %( 39<=v<= 42) 12387 6 2 1629 13 2 211 9 72 5 % 72 5 %( 43<=v<= 47) 14133 7 1 2223 15 7 253 5 79 9 % 79 9 %( 48<=v<= 54) 14741 7 4 2374 16 1 259 5 86 3 % 86 3 %( 55<=v<= 62) 12830 6 4 1528 11 9 191 9 92 9 % 92 9 %( 63<=v<= 71) 13245 6 6 753 5 7 91 6 100 0 % 100 0 %( 72<=v ) 14124 7 1 430 3 0 49 1 please refer to the “timiquickuserguide pdf” for more information about the charts displayed inside this summary report
