Na\u010d\u00edtanie d\u00e1t<\/h4>\nCarData: LOAD RowNo() as \"car_ID\", symboling, \"normalized-losses\" as \"normalizedlosses\", make, \"fuel-type\" as \"fueltype\", aspiration, \"num-of-doors\" as \"numofdoors\", \"body-style\" as \"bodystyle\", \"drive-wheels\" as \"drivewheels\", \"engine-location\" as \"enginelocation\", \"wheel-base\" as \"wheelbase\", \"length\", width, height, \"curb-weight\" as \"curbweight\", \"engine-type\" as \"enginetype\", \"num-of-cylinders\" as \"numofcylinders\", \"engine-size\" as \"enginesize\", \"fuel-system\" as fuelsystem, bore, stroke, \"compression-ratio\" as \"compressionratio\", horsepower, \"peak-rpm\" as \"peakrpm\", \"city-mpg\" as \"citympg\", \"highway-mpg\" as \"highwaympg\", price FROM [lib:\/\/Data\/Correlation matrix\/Automobile_data.csv] (txt, codepage is 28591, embedded labels, delimiter is ';', msq); <\/pre>\nSpustenie R skriptu<\/h4>\n
Teraz mus\u00edme pomocou R vypo\u010d\u00edta\u0165 korel\u00e1cie a pr\u00edslu\u0161n\u00e9 P-hodnoty. Nev\u00e1hajte pou\u017ei\u0165 nasleduj\u00faci skript. <\/p>\n
Tmp_Correlation: Load * Extension R.ScriptEval( '# install.packages(Hmisc, repos=\"http:\/\/cran.us.r-project.org\"); \/\/ponechajte # ak u\u017e m\u00e1te tento bal\u00edk nain\u0161talovan\u00fd library(Hmisc); df <- as.data.frame.list(q, strings.as.factors = FALSE); df <- Filter(is.numeric, df); cor <- as.data.frame.list(rcorr(as.matrix(df), type=\"pearson\")); cor$varname <- rownames(cor); cor;', CarData{wheelbase,length,width,height,curbweight,enginesize,bore,stroke,compressionratio,horsepower,peakrpm,citympg,highwaympg,price});<\/pre>\nTransform\u00e1cia v\u00fdsledku z R skriptu<\/h4>\n
R skript vracia kompaktn\u00fa tabu\u013eku obsahuj\u00facu v\u0161etko. Preferujem ju trochu vy\u010disti\u0165, aby sme skon\u010dili s pekn\u00fdm d\u00e1tov\u00fdm modelom. <\/p>\n
Tmp_Correlation_Crosstable: CROSSTABLE (Variable2, Value) LOAD varname as Variable1, * RESIDENT Tmp_Correlation ;\r\n\r\nTab_Correlation:\r\nLOAD\r\n,\r\nVariable1&'|'&Variable2 as %KeyCor;\r\nLOAD\r\nVariable1,\r\nReplace(Variable2, 'r.', '') as Variable2,\r\nValue as Correlation\r\nRESIDENT Tmp_Correlation_Crosstable\r\nWhere Variable2 like 'r.';\r\n\r\nTab_PValues:\r\nLOAD\r\n[P-Value],\r\nVariable1&'|'&Variable2 as %KeyCor;\r\nLOAD\r\nVariable1,\r\nReplace(Variable2, 'P.', '') as Variable2,\r\nValue as [P-Value]\r\nRESIDENT Tmp_Correlation_Crosstable\r\nWhere Variable2 like 'P.*';\r\n\r\nDrop Tables Tmp_Correlation, Tmp_Correlation_Crosstable;<\/pre>\nMali by ste skon\u010di\u0165 s dvojtabu\u013ekov\u00fdm d\u00e1tov\u00fdm modelom, ktor\u00fd vyzer\u00e1 takto:<\/p>\n
![\"D\u00e1tov\u00fd](\"https:\/\/emarkanalytics.com\/wp-content\/uploads\/2026\/04\/Datamodel.png\")
<\/p>\n
Vytvorenie korela\u010dnej matice v Qlik Sense<\/h2>\n
V\u00fdborn\u00e1 pr\u00e1ca! Sme pripraven\u00ed nastavi\u0165 vizualiz\u00e1ciu nad d\u00e1tov\u00fdm modelom, ktor\u00fd m\u00e1me. V\u00fdsledkom bude korela\u010dn\u00e1 matica zobrazuj\u00faca aktu\u00e1lne hodnoty Pearsonovho korela\u010dn\u00e9ho koeficientu spolu s farbami (tmavomodr\u00e1 pre negat\u00edvne korel\u00e1cie a \u010derven\u00e1 pre pozit\u00edvne). Okrem toho budeme ma\u0165 aj slider na v\u00fdber \u00farovn\u00ed P-hodn\u00f4t, ktor\u00e9 sa bud\u00fa zobrazova\u0165. T\u00fdmto sp\u00f4sobom m\u00f4\u017eete zobrazova\u0165 a analyzova\u0165 iba \u0161tatisticky v\u00fdznamn\u00e9 korel\u00e1cie a z\u00e1rove\u0148 si vybra\u0165 \u013eubovo\u013en\u00fa \u00farove\u0148 v\u00fdznamnosti. <\/p>\n
Vytvorenie korela\u010dnej matice pomocou heatmap grafu<\/h4>\n
Vyberte Heatmap chart z Qlik Visualization bundle a nastavte: <\/p>\n
\n- Data > Dimensions: <\/strong>nastavte 2 dimenzie Variable1 a Variable2<\/li>\n
- Vytvorte nov\u00fa premenn\u00fa vSign a ponechajte Definition<\/strong> pr\u00e1zdne<\/li>\n
- Data > Measures<\/strong>:
=Avg({<[P-Value] = {\"<=$(vSign)\"}>} Correlation)<\/code> <\/strong>; nastavte label na Pearsonov korela\u010dn\u00fd koeficient<\/li>\n- Appearance > Options:<\/strong> Nepou\u017e\u00edvajte mean in scale a namiesto toho pou\u017eite fixed scale. Nastavte Min Scale Value<\/strong> na -1; Max Scale Value<\/strong> na 1 a Minimum Horizontal Size<\/strong> na 0<\/li>\n
- Appearance > Design:<\/strong> Vyberte farebn\u00fa sch\u00e9mu Qlik Sense Diverging<\/li>\n
- Appearance > General: <\/strong>nastavte n\u00e1zov na Korela\u010dn\u00e1 matica<\/li>\n<\/ol>\n
Spravte heatmap responz\u00edvnu na \u00farove\u0148 v\u00fdznamnosti<\/h4>\n
Heatmap je nastaven\u00fd, ale st\u00e1le mus\u00edme prida\u0165 nejak\u00fd vstup na priradenie hodn\u00f4t do premennej vSign. Pou\u017eime na to slider.<\/p>\n
Spustenie R skriptu<\/h4>\n
Teraz mus\u00edme pomocou R vypo\u010d\u00edta\u0165 korel\u00e1cie a pr\u00edslu\u0161n\u00e9 P-hodnoty. Nev\u00e1hajte pou\u017ei\u0165 nasleduj\u00faci skript. <\/p>\n
Tmp_Correlation: Load * Extension R.ScriptEval( '# install.packages(Hmisc, repos=\"http:\/\/cran.us.r-project.org\"); \/\/ponechajte # ak u\u017e m\u00e1te tento bal\u00edk nain\u0161talovan\u00fd library(Hmisc); df <- as.data.frame.list(q, strings.as.factors = FALSE); df <- Filter(is.numeric, df); cor <- as.data.frame.list(rcorr(as.matrix(df), type=\"pearson\")); cor$varname <- rownames(cor); cor;', CarData{wheelbase,length,width,height,curbweight,enginesize,bore,stroke,compressionratio,horsepower,peakrpm,citympg,highwaympg,price});<\/pre>\nTransform\u00e1cia v\u00fdsledku z R skriptu<\/h4>\n
R skript vracia kompaktn\u00fa tabu\u013eku obsahuj\u00facu v\u0161etko. Preferujem ju trochu vy\u010disti\u0165, aby sme skon\u010dili s pekn\u00fdm d\u00e1tov\u00fdm modelom. <\/p>\n
Tmp_Correlation_Crosstable: CROSSTABLE (Variable2, Value) LOAD varname as Variable1, * RESIDENT Tmp_Correlation ;\r\n\r\nTab_Correlation:\r\nLOAD\r\n,\r\nVariable1&'|'&Variable2 as %KeyCor;\r\nLOAD\r\nVariable1,\r\nReplace(Variable2, 'r.', '') as Variable2,\r\nValue as Correlation\r\nRESIDENT Tmp_Correlation_Crosstable\r\nWhere Variable2 like 'r.';\r\n\r\nTab_PValues:\r\nLOAD\r\n[P-Value],\r\nVariable1&'|'&Variable2 as %KeyCor;\r\nLOAD\r\nVariable1,\r\nReplace(Variable2, 'P.', '') as Variable2,\r\nValue as [P-Value]\r\nRESIDENT Tmp_Correlation_Crosstable\r\nWhere Variable2 like 'P.*';\r\n\r\nDrop Tables Tmp_Correlation, Tmp_Correlation_Crosstable;<\/pre>\nMali by ste skon\u010di\u0165 s dvojtabu\u013ekov\u00fdm d\u00e1tov\u00fdm modelom, ktor\u00fd vyzer\u00e1 takto:<\/p>\n
<\/p>\n
Vytvorenie korela\u010dnej matice v Qlik Sense<\/h2>\n
V\u00fdborn\u00e1 pr\u00e1ca! Sme pripraven\u00ed nastavi\u0165 vizualiz\u00e1ciu nad d\u00e1tov\u00fdm modelom, ktor\u00fd m\u00e1me. V\u00fdsledkom bude korela\u010dn\u00e1 matica zobrazuj\u00faca aktu\u00e1lne hodnoty Pearsonovho korela\u010dn\u00e9ho koeficientu spolu s farbami (tmavomodr\u00e1 pre negat\u00edvne korel\u00e1cie a \u010derven\u00e1 pre pozit\u00edvne). Okrem toho budeme ma\u0165 aj slider na v\u00fdber \u00farovn\u00ed P-hodn\u00f4t, ktor\u00e9 sa bud\u00fa zobrazova\u0165. T\u00fdmto sp\u00f4sobom m\u00f4\u017eete zobrazova\u0165 a analyzova\u0165 iba \u0161tatisticky v\u00fdznamn\u00e9 korel\u00e1cie a z\u00e1rove\u0148 si vybra\u0165 \u013eubovo\u013en\u00fa \u00farove\u0148 v\u00fdznamnosti. <\/p>\n
Vytvorenie korela\u010dnej matice pomocou heatmap grafu<\/h4>\n
Vyberte Heatmap chart z Qlik Visualization bundle a nastavte: <\/p>\n
\n- Data > Dimensions: <\/strong>nastavte 2 dimenzie Variable1 a Variable2<\/li>\n
- Vytvorte nov\u00fa premenn\u00fa vSign a ponechajte Definition<\/strong> pr\u00e1zdne<\/li>\n
- Data > Measures<\/strong>:
=Avg({<[P-Value] = {\"<=$(vSign)\"}>} Correlation)<\/code> <\/strong>; nastavte label na Pearsonov korela\u010dn\u00fd koeficient<\/li>\n- Appearance > Options:<\/strong> Nepou\u017e\u00edvajte mean in scale a namiesto toho pou\u017eite fixed scale. Nastavte Min Scale Value<\/strong> na -1; Max Scale Value<\/strong> na 1 a Minimum Horizontal Size<\/strong> na 0<\/li>\n
- Appearance > Design:<\/strong> Vyberte farebn\u00fa sch\u00e9mu Qlik Sense Diverging<\/li>\n
- Appearance > General: <\/strong>nastavte n\u00e1zov na Korela\u010dn\u00e1 matica<\/li>\n<\/ol>\n
Spravte heatmap responz\u00edvnu na \u00farove\u0148 v\u00fdznamnosti<\/h4>\n
Heatmap je nastaven\u00fd, ale st\u00e1le mus\u00edme prida\u0165 nejak\u00fd vstup na priradenie hodn\u00f4t do premennej vSign. Pou\u017eime na to slider.<\/p>\n
![\"Korela\u010dn\u00e1](\"https:\/\/emarkanalytics.com\/wp-content\/uploads\/2026\/04\/CorrelationMatrix.gif\")
![\"Model](\"https:\/\/emarkanalytics.com\/wp-content\/uploads\/2026\/04\/modelPrice.png\")
<\/p>\n