This is an exemple feature selection. so without enrich your data set it very dificult to improve our score. Because all variable are very small contribution.
('Q1', 0.061429050143116733)
('Q2', 0.011806785042619903)
('Q3', 0.022095198979739418)
('Q4', 0.031424206677461951)
('Q5', 0.022997306223697462)
('Q6', 0.0075367235122939993)
('Q7', 0.063292254455874275)
('Q8_1', 0.0037811016699987399)
('Q8_2', 0.012228517973679081)
('Q8_3', 0.0019458127824748478)
('Q8_4', 0.012176487844446207)
('Q8_5', 0.00012773417517041364)
('Q8_6', 6.1050951915491139e-06)
('Q8_7', 3.0306989764214775e-05)
('Q8_8', 0.00074663998878069459)
('Q8_9', 0.0069981937618360975)
('Q8_10', 0.0080571753176474897)
('Q8_11', 3.1520189421350721e-06)
('Q9', 0.0035147251238743398)
('Q10', 0.020690137802578938)
('Q11', 0.0027867219352076316)
('Q12', 0.08675660716409786)
('Q13', 0.093381784338594009)
('Q14', 0.1089953085120534)
('Q15', 0.12481162722766649)
('Q16', 0.094188541036927476)
('Q17', 0.017972099566778598)
('Q18', 0.024868603998472433)
('Q19', 0.019999319770607104)
('Latitude', 0.06599559403683862)
('Longitude', 0.069356176833567679)
what methods are preferable for feature selection