Simultaneous Analysis of contingency tables drawn with telephone data registration from the National Telephone Service to Support Women Suffering Violence in Uruguay Elena Ganón Fundación Plenario de Mujeres del Uruguay (PLEMUU ) ganonelena@gmail.com International Conference on Correspondence Analysis and Related Methods CARME 2011 Agrocampus Ouest. Rennes. France. 8 -11 February 2011
OVERVIEW INTRODUCTION NATIONAL TELEPHONE SERVICE DATA DESCRIPTION METHODOLOGY APPLICATIONS CONCLUSIONS
INTRODUCTION National Telephone Service to Support Women Suffering Violence in Uruguay Created in October 1992 Organizations Local Government INTENDENCIA DE MONTEVIDEO Women NGO PLEMUU Telephonic Company ANTEL Multidisciplinary staff Statistical register
NATIONAL TELEPHONE SERVICE National Telephone Service to Support Women Suffering Violence Total phone calls 900 700 500 300 100 observed trend ‐ cycle
DATA Data registration: Victim call form Hour ( start, end ), date (day, month, year), operator code, phone-call origin ( Province ) Goal call ( information, advice, emergency) Women and aggressors characteristics: Sex, age, educational level, occupation Violence: Problem ( Domestic Violence / Non-Domestic Violence) Type (psychological, physical, economic, sexual, other) Aggressor Modality ( Treat / attack)
DATA- Transactional table llave_sertelmodelo_fichid_llamado fecha añol triml mesl dial diasemana hora_inicio hora_fin cod_operadora num_llamadonombr cod_demapor_qcod_objv_cod_sexov_edad 2 594 ene-2001 44 481 01/08/2001 2 001 3 8 1 4 10:50 11:00 18 1 1 3 1 62 2 595 ene-2001 44 482 01/08/2001 2 001 3 8 1 4 13:00 13:10 18 1 1 3 1 24 2 596 ene-2001 44 483 01/08/2001 2 001 3 8 1 4 13:30 13:45 35 1 silvia 1 3 1 39 2 597 ene-2001 44 487 01/08/2001 2 001 3 8 1 4 21:22 21:30 15 1 GLADI 1 3 1 43 2 598 ene-2001 44 488 01/08/2001 2 001 3 8 1 4 23:00 23:06 15 3 JAQUE 1 3 1 999 2 599 ene-2001 44 492 01/08/2001 2 001 3 8 1 4 10:08 10:14 19 1 NANCY 1 3 1 37 2 600 ene-2001 44 497 02/08/2001 2 001 3 8 2 5 13:30 13:40 16 1 Maria 1 3 1 43 2 601 ene-2001 44 498 02/08/2001 2 001 3 8 2 5 14:25 14:40 16 2 Maria d 1 3 1 999 2 603 ene-2001 44 501 02/08/2001 2 001 3 8 2 5 18:00 18:10 16 1 Jaquel 1 3 1 33 2 604 ene-2001 44 504 03/08/2001 2 001 3 8 3 6 09:53 10:01 19 1 1 3 1 999 2 605 ene-2001 44 507 03/08/2001 2 001 3 8 3 6 10:44 10:58 19 1 Gabrie 1 3 1 26 2 606 ene-2001 44 508 03/08/2001 2 001 3 8 3 6 11:21 11:27 19 1 Alejand 1 3 1 33 2 607 ene-2001 44 509 03/08/2001 2 001 3 8 3 6 11:29 11:31 19 1 Gabrie 1 2 1 21 2 608 ene-2001 44 511 03/08/2001 2 001 3 8 3 6 12:06 12:12 19 1 Silvia 1 3 1 34 2 609 ene-2001 44 512 03/08/2001 2 001 3 8 3 6 13:38 14:00 16 2 Maria B 1 3 1 28 2 610 ene-2001 44 513 03/08/2001 2 001 3 8 3 6 16:40 16:50 16 1 Maria 1 3 1 45 2 611 ene-2001 44 514 03/08/2001 2 001 3 8 3 6 20:50 21:05 21 1 Olga 1 3 1 68 2 612 ene-2001 44 515 03/08/2001 2 001 3 8 3 6 21:10 21:20 21 1 Ma del 1 3 1 38 2 613 ene-2001 44 516 03/08/2001 2 001 3 8 3 6 22:30 22:40 21 1 Maria 1 3 1 44 2 614 ene-2001 44 517 03/08/2001 2 001 3 8 3 6 22:50 23:05 21 1 Magela 1 3 1 24 2 615 ene-2001 44 518 04/08/2001 2 001 3 8 4 7 09:28 09:43 19 1 Blanca 1 3 1 70
METHODOLOGY – INTRODUCTION DATA EXAMPLE A7aAlt30 A7aA3039 A7aA4049 A7aA5059 A7aA6069 A7aAgt69 A8aAlt30 A8aA3039 A8aA4049 A8aA5059 A8aA6069 A8aAgt69 A9aAlt30 A9aA3039 A9aA4049 A9aA5059 A9aA6069 A9aAgt69 Vagelt30 289 178 39 16 8 3 335 193 55 34 3 4 333 209 57 27 11 2 V30age39 61 358 212 46 17 2 86 428 214 64 20 7 67 388 238 46 32 6 V40age49 65 53 265 163 30 11 69 56 309 154 30 11 78 55 291 146 30 15 V50age59 64 19 52 128 80 22 68 37 34 171 67 17 68 48 33 168 57 28 V60age69 24 23 21 21 63 29 18 30 13 23 51 30 31 35 16 20 61 32 Vagegt69 6 8 15 12 6 40 9 8 17 13 7 26 7 13 26 12 12 42 CA separate table Table total A7 2449 A8 2711 f i j = a i j /a i=1,6 j=1,6 f 12 = 178/2449 A9 2740 stacked table 7900 CA juxtaposed table SA concatenated table f i j k =a i j k / a . . k f 121 = 178/2449 f i j =a i j /a i=1, j=18 f 12 =178/7900
METHODOLOGY - CA Correspondence Analysis (CA) Absolute frequencies matrix A A = (a i j ) i=1,..,I; j=1,…,J a = i j a i j f i j = a i j /a Relative frequencies matrix F F= (f i j ) Column profile {f i j / f . j , i =1,.. I } Distance between column profiles j , j’ d 2 (j , j’) d 2 (j, j’) = 1/f i. (f i j / f .j - f i j’ /f .j’ ) 2 Calculation of eigenvalues and eigenvectors of X’X , X = (x i j ) x i j = √ f i. (f i j / f i. f .j – 1) √ f .j * Juxtaposed Tables Absolute frequencies matrix A A = (a i j k ) i=1,..,I; j=1,…,J k , k=1,...K a = i j k a i j k Relative frequencies matrix F F= (f i j k ) f i j k = a i j k / a Simultaneous Analysis (SA) Absolute frequencies matrix A A = (a i j k ) i=1,..,I; j=1,…,J k , k=1,...K a .. k = i j a i j k Relative frequencies matrix F F= (f i j k ) f i j k = a i j k / a ..k
METHODOLOGY -SA Simultaneous Analysis (Zárraga & Goytisolo,2002,2003) Absolute frequencies matrix A A = (a i j k , i=1,I, j=1,J, k=1,K) a . . k = i j a i j k Relative frequencies matrix F F = ( f i j k , i=1,I, j=1,J, k=1,K) f i j k = a i j k / a . . k Column profiles {f i j k /f . j k i = 1,..,I} Distance between columns profiles d 2 (j , j’) d 2 (j, j’) = 1/f i . k (f i j k / f . j k - f i j’ k /f . j’ k ) 2 Calculation of eigenvalues and eigenvectors of X’X , X = (x i j ) x i j = √ k √ f i . k (f i j k / f i . k f . J k – 1) √ f . j k Where k could be 1 , or 1/ k being k the first eigenvalue of table k, or 1/ k reciprocal of total inertia table k Projections on axes: overall rows and columns, overall and partial rows, tables relation between factors of separate CA and SA
IMPLEMENTATION Original data: transactional table (spreadsheet, relational data base, OLAP cube) individual calls * variables Generation of table for analysis contingency table or juxtaposed tables Use SimultAnR R-program by Zárraga&Goytisolo (2010) Load the R base executable module Load the R application executable module Read the data table with read.data command Use SimultAnR commands to perform the analysis
IMPLEMENTATION - commands copy to the working directory the .txt file run the R-program load working directory load SimultAnR comands datoleido_2 <- read.table("TAVpfnd_2a5.txt",header=TRUE) read data datoleido_2 print on the scrren the data dataSA<-datoleido_2 assign the data table SimAn.out <- SimAn(data=dataSA,G=4,acg=list(1:3,4:6,7:9,10:12),weight=2,nameg=c("2","3","4","5")) made the analysis SimAnSummary(SimAn.out) shows the results at the screen SimAnGraph(SimAn.out) shows the graphics at tha scrren SimAnGraph(SimAn.out) generate a .pdf file with graphics pdf('SAGr.pdf',paper="a4r",width=12,height=9) SimAnGraph(SimAn.out,s1=1,s2=2,screen=FALSE) dev.off()
APPLICATIONS CA of tables violence problem * years_2002_2009 violence problem * (age) * (education) * (occupation) women victim age * years 2002_2009 women victim age* occupation Juxtaposed tables age * occupation * violence problem age * education * violence problem age* occupation * years
APPLICATIONS 2a age * education * violence problem Simultaneous Analysis age*education * violence problem DP_prim DP_sec DP_higher DR_prim DR_sec DR_higher ND_prim ND_sec ND_higher 2068 2618 302 178 353 79 24 46 9 agVvlt 30 2227 2945 728 208 237 55 47 43 12 agV30a39 1759 2115 599 280 363 95 47 49 15 agV40a49 852 853 251 305 279 96 30 28 9 AgVv50a59 418 217 65 274 129 43 40 14 6 agV60a69 184 57 11 211 61 12 24 7 4 agVgt69 separate CA inertia percentage table axis eigenvalpercentage cumulated CA 58.8 58.8 DP 1 0.0135 41.2 2 0.0094 100.0 rows and columns 96.4 96.4 DR 1 0.0698 3.6 2 0.0026 100.0 projections 98.2 98.2 ND 1 0.0618 1.8 2 0.0012 100.0
APPLICATION 2b age * education * violence problem Simultaneous Analysis SA juxtaposed table SA overall rows and columns overall and partial rows SA tables projections SA Projections of tables Contributions of tables to SA table Axis 1 Axis 2 Axis 1 Axis 2 tables projections 0.82 0.79 0.3 0.96 DP factors relation ca & sa 0.96 0.35 DR 0.01 0.01 0.96 0.35 ND 0.03 0.04
SA - age * education * violence problem
age * education * violence problem
Recommend
More recommend
