gis reports of part one crime numbers are telling a different story of crime in the city. when they should be the same. the comptroller's office found that the primary driver of these differences in part one crime numbers is the fact that the method by which the department counts crime for the profile difference in the method by which they can crime for the other report. these methods should be aligned. in fact, the difference in reported aggravated assault is the primary reason why one line is higher than the other line. specifically, the department is under reporting domestic violence aggravated assault to the department of justice because the methodology for counting aggravated assaults in the ucr reports does not include domestic violence cases. commissioner: i am sorry.

can i just ask you a question before you go on? >> yes. thank you. commissioner: i understand that we would hope and expect that the lines from what we report are the same, but the department of justice defines of violent crime differently. is it not by definition going to have to be different? >> we looked into this and compared the data, and the definitions are, indeed, the same. there are characteristics that make these numbers slightly different, but they should not account for the disparity that you see in this graph. commissioner: ok, i thought you just said they do not recognize domestic violence as a part 1 crime while we do. did i miss here you? >> no, they do.

aggravated assaults, that should be accounted for in the aggravated assault number that is reported to the department of justice through the ucr report. commissioner: and we are not reporting it because -- >> the methodology does not include domestic violence /aggravated assault, where is the profile does. that is why you have that disparity. commissioner: sorry to interrupt. they count it as an aggravated assault, but we consider it differently. commissioner: ok, i am sorry. what you just said does not necessarily account -- there are two things that on.

one is under reporting, but that graph shows more than under reporting. in the box that you called out, a complete divergence. that portion of the graphs generally shows the crime trend and help it is exactly going to gather. there is one point where we show crime going down. we are locally reporting to the department justice that it is going up. do we have more for that one portion, or is that what you just explained to me? >> the findings that we just explained are counting for the divergence. commissioner: so i cannot discern -- >> march 2011. we are showing these are different between these two

ports, and in addition, the trend in crime is also different, which is shown in the orange college, and there are a variety reasons of why that should be different, and this is the finding that the comptroller's office had, where there is not a single system of record, and there are multiple, a disparate data sources in the way they report crime. commissioner: i know you are going to go on, and i apologize be intense about this, and i know i am probably being simplistic about this, if it is going into one, why is it not going into the other report the same way? >> the disparate data sources, there are different data sources that populate these two reports. commissioner: i got you. >> any other questions?

>> of the questions you were asking what prompted this report. >> sorry. president mazzucco: and now we have the solution. >> many of the findings i just discussed will be worked on with what is currently in development. however, the comptroller's office identified areas for immediate improvement. but first as having the homicide rate and shooting logs directly interfaced with the database. we believe this would eliminate many of this manual processes that i was talking about that introduce errors into the profile. secondly, we are recommending automatically generating the profile directly from the database. this would also eliminate many of the human errors that are made to date, currently the process by which the department

develops this profile is very manual, so we are suggesting that it be automated, so we would avoid many of those manual processes. the third recommendation is using the same methods to account crime in the reports. there are different methods by which they are counting crime separately for these reports, when they should be the same, and if they were the same, it would provide a more consistent public report of crime in san francisco. and lastly, we are recommending delaying the delay of the profile by at least four days after the last day in the reporting period. we think this would minimize the impact of system data delays on crime comparisons. we think that these short-term recommendations could significantly improve the accuracy and consistency of reporting by the department.

>> oh in addition, the comptroller's office has identified recommendations for the crime data warehouse to have accurate, timely, and consistent crime reporting in the long term. for example, in our memo, we have identified key functions here -- features that should be part of the system. we have also talked about the need to enforce and increase training on standard report writing to assure accurate crime reporting by the department. so next, i would like to hand off the discussion of the crime data warehouse highlights to the information officer for the police department. president mazzucco: good evening, miss. how are you? >> good evening, commissioners. most of you probably do not know

what the crime data warehouse is. it is a massive undertaking. under the leadership of chief suhr to have a data warehouse that will be the single one and only plays where we have crime did in san francisco. we have completed the two phases of this project. the crime data warehouse will be a portal of all 24 plus criminal justice systems that we use today, that police officers use to prevent, manage, solve crime. they will be in one location, one user id, 1 password, highly graphical, it's easy to use, etc., and we have completed two phases, and one is the search capability, and the second that we are in right now that we just launched january 17, the police officers are entering crime

reports directly into the database, so we are essentially moving from primarily a paper base system to an automated system, so how this is relative, if you can imagine, this piece of it is really just counting and categorizing crime. it should be fairly simple, but because of the many years and many layers and many processes and the many sources, all of the different units can crime, and they are all trying to really stay on top of this, but as a result, we had discussions. i do not think we can explain all of the differences. what the warehouse will allow us to do, and again, we are live on this, 200 people, and we are rolling it out right now, we have gone past actually the most difficult part of the project. i can say with a great deal of confidence that this is happening. it will be a single source for

our crime data, so we will have one and only one place where we go to count crime. the crime data will be available immediately, so the minute the officer -- the minute the incident report is approved, it will be available for counting, searching, so that is at most a 10-hour window as opposed to the two, three, five or even longer window that we have before a crime can be viewed or counted. and everyone will have access to the same sort of data, so if a crime is ms. categorized, and if i am the captain of the tenderloin, and i am seeing 25 burglaries on my report, but i see 25, where it should be a robbery or not a burglary, i will go into the crime data warehouse, the same place where the data is, and i will go in

and we classify it purpose -- -- reclassify it purposefully. it is a radical simplification of this process. and then finally, the crime did a warehouse, it is about data and accountability. the crime data warehouse will allow us and much improved opportunity for accountability because we have this data. we can sort it at the district station, geography, officer. we can sort it any number of ways and improve our accountability which will give us visibility in processing things that are going wrong, so one of the first things we saw is that some of the stations, a crime that was committed over here was actually getting reported over there. we have not had this ability in that. so a new opportunity, i think this problem will be greatly

minimized. it will never go away. you always have issues. we have a great opportunity to improve this process. >> we will be able to go back and reclassified the crime. -- president mazzucco: we will be able to reclassify a crime. when i heard about is that if somebody it tends to break into my house, is it an attempted burglary, or is it a vandalism? or the classic scenario where someone steal something from a gap store, and they knock down a clear, is in a robbery, or is in effect? when is that changed and put in? >> oh right now, the way we do that is a green sheet. the inspectors go downstairs. they talk to whoever, who the case is, and they decide whether

it was to keep the property or if it was in fact an incidental and knocked down and you were for being the place, where one would be grand theft, and the other would be a robbery. the vandalism versus the burglary, it could have been the person we just had on the ledge on telegraph hill. we do not know if he ever went in. we just know that he broke a window, so these things are all done now manually on a piece of paper. what susan is saying, or this person says, no, i disagree," once the data warehouse is on- line, it will be a push of a button, and then the end code will also change, which will make our reporting more accurate going forward for the guidelines. president mazzucco: thank you.

>> what they have done, and i cannot think the comptroller enough, what we have said before when we repair -- prepare our police reports electronically, that just means that the word processor, the typewriter we use, is plugged in. that does not mean that that data has gone anywhere, so right now, we are collecting data, and we are back to 2008 on old data. >> the old did it goes back to 2000. different dates for different things, but different dates. >> and we have made cases with this new system, and going forward, we will implement the recommendations, so we will implement these recommendations. we will get on track. our information will be reliable, and it will only get better as the data warehouse goes forward, and then the vision, if you will, is to connect with the other counties to get their data available to

us, too, online and in the computers on the web, so at the end of the day, we will be faster, more efficient, and cheaper in the long run than we ever been, and the accuracy will go without saying because the manual will all go away other than the police work that needs to be done to generate the statistics in the first place. president mazzucco: commissioner turman? commissioner turman: i had some issues with the recommendations. number three, it should use the same method of counting crimes for the crime report. what is the best way? how can we do that, especially when we have such different reporting in different stations

as well as the department of justice? what is your recommendation as to how to get that accomplished >> right now, we use codes to crowds -- count crime. a homicide will have various codes that are used to count the number of homicides reported by the department. and so the doj report also is calculated by the department using these codes, said they will count how many codes they have in the reports received, and the same is true for the profile, so what we noticed is we analyze the codes, and that is the method by which these two reports are counting, and we noticed that there are differences in how these codes are being used by crime type, so for the example of aggravated assault, we noticed the codes that associated aggravated

assault cases were being used for one profile but were not being used to count of aggravated assaults in the other report. so it is really that we are continuing to rely -- if the department is continuing to rely on using these codes as their methodology by which crime is accounted for these two reports, those codes should be the exact same way for both reports. does that answer your question? commissioner turnan: i understand how we're coming up with different ones, but my question was how to get everyone to come together with one methodology. that is what i am talking about. has that dialogue begun? how do we do that is my real question.

>> right now, some of the numbers are aggravated from three spots, and we will cut that down to a maximum of two, which gives us a 33% better chance of counting it right the first time, and for the rate in shootings, and i believe we added db to that, right? we are going to counter those individually prove we're going to take the mystery out of them, and we will have a single source for the most serious crimes, so we will get that right. there will be a single source. commissioner turman: does that match the way they are doing it? >> we are committed to getting it right, and this is a fabulous first start. commissioner turman: thank you. president mazzucco: commissioner

kingsley? commissioner kingsley: thank you for getting the ball rolling. sounds like exactly the right thing and much needed. thank you for your hard work on this and reporting back to was an explaining what is happening here. it sounds like this is just right, spot on, in terms of going in the right direction. could you please give me a better understanding of how this works in a kind of day-to-day way for the police officer on the street? to the data or the crime actually getting into the warehouse, how that process is going to be different from what is happening now. one question that comes to mind, is there an electronic toll that the officers will have at the site of a crime?

are they still going to be writing notes down on paper and taking it back to a station? how is that process point to change or remain the same? >> ok. well, today, it is hard to believe, but the officers actually type the police report. they print it. this is today, prior to ingleside. 2012, right. they tied in the police report. they print it out. somebody takes a stack of this paper reports to room 400 at the hall of justice. some beleaguered people on a 24/7 shipped type that information, most of it, not all of it, into an old, 40-year-old system, which contains some pieces of data but not all, for example the narrative it does not contain, which is one of the most important things.

that is a description of what the crime really was, and then that data is used to marry the local version of what was contained on that desk top. we would be here all night to really explain all of the pieces and parts of this puzzle. there is another component which is citizens entering police reports online, and that plays into this, as well. how it will be in the future, and what is happening right now at ingleside, the officers are entering their police report directly into the data warehouse. in the future, it will go nowhere. currently, because we have to run simultaneous systems, we have the old, and we still have to keep this machine running, until we have fully rolled out on the new, we have had to keep the old process in place, so they are typing the reports into the new system. they are fully available online. they are doing electronic

signatures on line, which is a completely new process, but they are still coming just for now, printing the report and going through the old process. the reason for that is we have 40 plus agencies who all use the back processes to do their day- to-day jobs, so we could not cut that off with the new system. it is a long story, but in the future, the police officer enters the report directly into the crime data warehouse. that is it. notepaper. it does not ever have to be printed out at all, ever. commissioner kingsley: this is being used by computers with a regular software. >> with a computer, you can access the system. commissioner kingsley: how is

this going to be different? >> the system has never been like a computer system. it has been several different computer systems and a lot of people a aggregating information to populate a table, so in the future, the data warehouse will be the computer and will populate whatever we want from the data warehouse in almost real-time because the information will go right into the data warehouse and immediately become available and searchable. to populate reports, to report to the commission if you are curious about something, four detectives it's a gold car was used, and the guy has a nickname of "mookie" and it will look up every "mookie," but right now,

people will anecdotally have to remember. it is in my locker, or i am off today. angela, go in my locker and get my note pad. commissioner kingsley: is this replacing this? >> this comes off of an excel spreadsheet. i do not know if you know what that is. that will be no more. there will be no more 20 plus sources of data coming into the excellus spreadsheet. there will be no more dialing for dollars, mine is different from yours, all of that process, because once the officer enters that incident report, that is it. that is the crime. that is the count. it is wrong or if they miss classified it, the owner of that crime, who would be the captain of the district station where the crime occurred, would have that in dr. kapoor, said they

will get a report that says here are all of the crimes. it is a really simple system, really easy to see what comes into your account. -- count. if they see something is wrong, they can we classify it right then and there. the difference from the spreadsheet, i do not know where this came from, it came from all of these different sources, and i do not understand it. >> that is just a phrase new york came up with that means computer-generated statistics. we are finally going to have truly computer-generated statistics, and we will keep the name comstat. we will for the first time ever in san francisco have truly computer-generated crime statistics in near real-time,

and in the meantime, the comptroller's office has given a mechanism to us to put a band- aid on the process to make it better than it has been until we get to be searchable data warehouse. commissioner kingsley: well, it sounds terrific. are we ahead of the curb, or are we behind the curve -- are we ahead of the curve? >> we will be way ahead of the curve. >> sometimes it is good to start off from behind, because we can start from scratch, which is what we did, and is just unbelievable how we are moving forward when we languished literally for better than a decade or more trying to get to where we hopefully maybe by the end of the year. >> yes, the new system. one of the things that has is a google-type search ability.

if a police officer enters something, a nickname, the description of a tattoo, and physical description, anywhere, and then later that same set of words is search, it will pop up with the incident where that has occurred, so it is really advanced, even more than the average law-enforcement system search capability, and it is very easy to use. it is very graphical, very simple. it is pictures, photographs, and it is great. commissioner kingsley: wonderful. one last question, if i could. if anyone can go on with the password, how is there either control or knowing who changed or him made an entry or changed and entry, etc.?

a tracing, basically, on where the information came from? there might be an incident with 12 people in it. how do we know who is reporting what? >> the system has an audit. anytime a report gets touched or changed, it creates a new version that shows exactly who changed what and when and their star number, etc., and the system will also be locked down within the sfpd internet, and in terms of sharing the date outside of sfpd, which we have done, public information, this creates a new set of challenges for us, because we handle security by paper because you have to go to the records room. this is a challenge, and we are working on how that is going to work. commissioner kingsley: thank you. president mazzucco: how many

officers will this put out? how many more personnel will be available? >> certainly, it will put some of the officer's back on the street, but the greater benefit is once, but some type of electronic tablet, it is the amount of time that will keep officers back on the street because they will not have to come back to the station to use the desktop computers at the station. they will be staying outside, that half-hour trip back and forth that they make three or four times, it should increase their time on patrol by about 30% per officer. president mazzucco: that is incredible. that is a question we had before, so thank you for building the system. >> and thank you for the controller's office for getting it right. president mazzucco: