Frequently Asked Questions


What is it that you do?

We support communities and law enforcement and assist them in improving the quality of life in a community.  We do this by helping to solve crimes that would otherwise go unsolved. Currently the case clearance rate - the number of cases that are solved - is very low.  Typically in major cities, 14% of property crimes and 30% of violent crimes are ever solved.  That means that 86% of property crimes and 70% of violent crimes are not solved.  The primary reason for this low closure rate is usually due to a lack of witnesses and leads to solving the case.  We provide what is essentially a persistent witness to events. 


How is it that you reduce crime?

We work to reduce the crime rate in a city in three ways -  increased deterrence, removal of repeat offenders early, and pushing crime out of the area. 


1. Increased Deterrence - The number one way we hope to decrease crime in the areas we support is by increasing the deterrence for people to commit crimes.  The number one factor in deterrence is the perceived likelihood of getting caught and convicted.  We and those we support would always rather deter a crime than have to solve it.  By increasing the likelihood of the criminal getting caught,  we hope to prevent the crime altogether.  


2. Removal of Repeat Offenders -  There is a very small percentage of people in a community that commits a very large portion of the crimes.  The average offender will commit 18 or more crimes before getting caught.  Many will commit hundreds of crimes.  By increasing case clearance rates and removing career criminals sooner, communities become safer. 


3. Push Crime Out of the Area - The third way we lower crime in areas we support is to push it out of the area.  If you stand a significantly higher chance of getting caught in an area with cameras then a criminal might just move to another area without cameras.  Baltimore areas with CitiWatch cameras have seen a significant decrease in crime to the point many community groups are asking for cameras to be placed in their areas.   We provide the equivalent to hundreds of cameras over a major part of a city at once. 


How Does Deterrence Work?

Deterrence works by having people not commit crimes due to the potential punishment they might receive.  The major factors of deterrence are the likelihood of getting caught, the likelihood of getting convicted, and the amount of the punishment.  The likelihood of getting caught is the primary factor in deterrence. 


We know that when Baltimore put in a lot of ground cameras crime dropped significantly in the view of those cameras.  Because of this effect, many communities are now asking for additional cameras.  Deterrence requires people to know of the capability and the increased likelihood of getting caught.  We believe that by being open about the communities enhanced ability to solve crimes accompanied by an increase in closure rates and convictions will dissuade people from committing crimes knowing that they stand a much higher chance of getting caught. 


Is it legal?

Airborne imagery has been used in law enforcement for more than 50 years.  Many police agencies currently use helicopters to assist officers and detectives to solve crimes.  The CSP program is no different.  There are 4 Supreme Court level decisions that outline the law associated with aerial surveillance.  In each case, the use of airborne law enforcement has been upheld to be legal and admissible in court. The CSP provides a memorandum of law that addresses these cases and it is available on our website.


Can it be used in court?

Yes, the information we provide can be used in court.  The information we provide is just one small part of court proceedings and it is combined with witnesses, forensic evidence, and other information to support search warrants and trials.


Do you follow everyone or just those involved in crime?

We only follow people who are at or near a reported crime scene.  Within our 32 square mile images, there are likely more than a million people. We could not possibly follow everyone nor would there be any reason too.  During a typical flight, we will receive hundreds of call for service where people have reported crimes and asked for help.  We prioritize these calls by their severity and first work most serious.   In a major crime, such as a murder or shooting, we will follow all the cars coming and going from the crime scene.  This takes time and requires an analyst to follow the cars.  Through this process, we identify where the people at the crime scene went to and came from.  We cannot determine specifically who did it or who might be a witness and who might be an accomplice. We can only determine that a car came from this location to the crime scene, and after the crime when to this location.  We provide the information to detectives who use the information we provide to assist them in their investigations. 


What is it that you provide to investigators?

We provide a high-level view of the crime scene, the cars and the number of people who were there, where those cars came from and where they went to, and their actions while going to and from the crime scene.  In a recent shooting, we followed 6 cars to and from the scene.  Four left the scene within seconds after the shooting.  Some drove erratically running both stop signs and red lights as they fled the scene.  From following those cars, we were able to identify 8 houses that the cars either came from or went to before and after the shooting.   This information was passed to investigators.  In that case, the victim was not willing to assist investigators.


If you cannot identify people what good is it and how is it used in court?

While we cannot identify people we can track them to and from crime scenes.  Tracking a person from a crime scene to a car and then tracking a car to an address provides useful information to an investigator.  We can also provide an overview of the number of cars and people at the scene.  This allows corroboration of witness statements and other information from the crime scene.  Once you have an address, officers can go to that address and identify specific vehicles and obtain a license plate number, they can search the police databases for previous information on the address.  As we track a car it often passes several ground-based cameras allowing an identification of the make and model and often a license plate.  These images can be useful in aiding a detective in solving a crime.  


What happens when your cameras get better?

One of the people's fears is that eventually the cameras will get better and people will be able to be identified by our cameras.  This would take a massive improvement in all aspects of our systems.  We currently use a 192-megapixel camera to capture a 32 square mile area.   If we had a 10 fold improvement in our cameras we would have 10 pixels on a person.  At 10 pixels on a person, you still cannot tell anything about a person.  At the same time, we would be covering 1/10 the amount of area that we could cover at 1 pixel per person and see 1/10 the amount of crime.   From an operational perspective, it is better to watch a larger area and see more crime than it is to have more pixels on a person. 


Do you follow people from protests?

We do not follow people from protests.  Our privacy policy is very clear that we will not follow people involved in peaceful protests.  If protests become a riot and crimes are committed we would be authorized to follow the people and cars suspected of being involved in crimes to support investigations. 


What privacy protections do you have?

CSP has developed a strict and comprehensive privacy policy that controls and limits what we can look at and the information we can provide. It is part of our contract with our supported cities.  It lays out what we can support and what we cannot.  It was developed with input from a wide range of sources including local police departments, local community groups, State and National ACLU, and many others.  We have presented our program and policies at the Headquarters of the National ACLU and at some of their sponsored conferences.  We have answered all of their questions and have been open about what we do and how we do it.  Our privacy policy that resulted from community input and discussion is included on our website.  All of our analysts are trained on the privacy policy, what it allows and what it does not allow.  After being trained each analyst signs the policy and agrees to abide by it.  Our privacy policy is also placed on contract with our customers to ensure they understand it and that we adhere to it.  


Can you tell if people look at places where they are not supposed to?

Our imagery is served by a central server and it records where each analyst has looked over time.  This information is reviewable by management and outside oversight if and as necessary.  All tracks of vehicles are recorded and are associated with a specific crime or investigation event.  All tracks are reviewed by a senior analyst as the investigation briefing is compiled prior to briefing detectives. 


Is tracking of people and cars automatic?

No. People at crime scenes are tracked by analysts who visually follow the cars second by second and put track points on them.  CSP does not use computer automated tracking for a wide variety of reasons. these auto tracking systems simply do not consistently provide the accuracy needed for our programs.  The US Air Force and Army have spent hundreds of millions of dollars working on automated tracking programs using similar data.  They usually form "tracklets" of 20 seconds or so of a car before losing it and then they use statistics to determine which other "tracklet" the car might be.   This is just not good enough for use in legal analysis.  CSP uses people to review the data and to track suspect cars.  We find that people are much better at tracking cars than computers.  While this takes more people it ensures accuracy.  We typically follow a small number of cars per day that are directly involved in crimes within our imagery. 


How long do you keep the data?

We keep our imagery data for 45 days unless there are ongoing investigations or prosecutions associated with the data.  In the case where there are investigations and prosecutions, the data is removed from our active server and stored on disk drives in classified safes to protect it and save it as evidence.   Once it is removed from our server it is not accessible to our analysts without a specific request.


Can someone be convicted using just the data you provide?

It is very unlikely someone would be convicted off of our data alone.  We provide leads where often few are available.  For example, if we follow a person from a shooting to an address.  We can identify people who are associated with that address and match that with witness statements and other available ground-based videos.  If there is sufficient evidence a search warrant could be obtained.  In the house, you may find the gun, the clothes with gunpowder residue, and hopefully, the people involved.  It is likely the strong forensic evidence that will lead to the conviction or the plea bargain, not the fact that the person was followed to the house.


How many crimes do you see?

In a typical day, we see over 300 calls for service within our coverage area.  These often include 20 to 30 priority 1 calls including aggravated assaults, shootings, stabbings, and fires.  We commonly witness more than 30 priority 2 calls including robberies, hit and runs, common assaults, armed persons, destruction of property.  We commonly witness 60 plus priority 3 calls including stolen autos, burglaries, accidents, alarms.  We will also witness 100 priority 4 calls for narcotic sales, vehicle disturbances, common assaults, and injured persons. 


We cannot see all the crime that occurs within our imagery and do not have time to investigate all the events we can see.  We focus on the most serious crimes first and work our way down in importance as time allows.   A significant investigation with many vehicles and long tracks can take 3-4 analysts several days to complete.


Are you flying 24/7?

We typically fly 100 to 150 hours per month over an area.  We find this is easily done with a single aircraft and systems.  We could fly more but it would cost more.  We fly during peak periods of crime and over those areas as found by analyzing the crime statistics for a city. With the number of hours we currently fly we cannot analyze all the data and crimes we observe.


Are you targeting certain neighborhoods?

We do not target neighborhoods and we do not look at locations trying to predict crime.  We respond to calls for service and reported crimes where people have asked for assistance.  Our coverage areas in Baltimore cover 5.8 x 5.8-mile areas or up to 32 square miles.   When we are flying our Western orbit that includes everything from Johns Hopkins Medical Center in the East to Edmonson Village on the West.  We see from Druid Park and Johns Hopkins University to the North and Port Covington to the South.  We do not target any neighborhood but support all of them at once.   


How do you determine where to fly?

Where we fly is determined by our customers.  We typically base our flights on recent crime statistics to try and maximize the number of people we can support.   


How is the data secured and can someone get into it?

Our systems are secured highly secured through a variety of systems. First, they are maintained in a secure location.  We have cameras and other security systems that ensure that only authorized people have access to the imagery. Once in the system, we record where the analysts have looked and what investigations they have worked on.   All tracks are recorded on the server and are reviewable by the senior analysts to ensure that they are associated with the investigation the analyst was assigned.  The backup data is also maintained on disk drives stored in secure safes.      


Who has access to the data?

CSP analysts are the only ones with access to the imagery data.  CSP analysts are trained on how to analyze the data and evaluate what they see.  CSP analysts undergo a significant training program and sign commitments to follow the CSP Privacy Protection Program. 


Who looks at the data?

CSP analysts analyze the data.  Police and others are shown the results and the information associated with the investigations.  Police do not do the primary analysis but do look at the results of the analysis and the resulting tracks.   CSP hires local analysts and provides training to them on the systems’ methods and techniques to analyze the data. 


What is done with the data?

The information we develop from our imagery is briefed to detectives, prosecutors, defense attorneys and the courts.  It is included in the investigation files of the detectives.  It is presented to prosecutors in preparation for filing charges.  It is provided to defense attorneys as part of discovery. It is presented when needed to support trials. 


Are the Analysts trained?

CSP has training classes to teach our analysts the skills required.  These classes are up to 80 hours in length and have been incorporated into community college courses.  Skills we teach are sometimes specific to our systems but also include the investigative process, the role of law enforcement, the rules of evidence, we also teach systems specific skills such as video analysis, criminal techniques, and procedures we have observed.  We also teach skills such as public speaking and briefings and power point.


How do I become an analyst?

CSP continues to hire people who meet our requirements.  CSP hires qualified candidates who are honest people with computer skills and can present a professional image. 


Who would present the information in court?

CSP analysts and experts would present the information to investigators, prosecutors, defense attorneys and when needed the court.


Do you ever skip cars and follow the wrong car? 

Our analysts are very good at tracking cars through even difficult situation.  There are times when it is very difficult to determine which car emerges from behind a building.  An example, if a cars go behind a building and two cars come out and our analysts cannot determine which car is the suspect, they will track both cars and include comments in the notes and in the briefing of what they did and saw. Other one or both cars will pass a ground based camera and be verified as the suspect car and that track will be continued.


How long can you follow a car?

We can follow a car as long as it is within our imagery.  In several cases in Mexico, we have followed cars from murders for over 5 hours.  Through that track, we were able to identify 12 specific locations which aided in taking down a significant part of that group quickly. 



How many murders have you seen and how many have you solved?

To date, we have witnessed 36 murders as they occurred.  The information we have provided has led to the confessions that account for 75 murders so far.   Some of the murders have been fairly simple some have been quite complicated multiple car events.  Our systems allow multiple cars to be tracked from a crime scene at once to help stop all of the people involved. 


What is the lowest crime you have seen and investigated?

The lowest crime we have investigated was a necklace snatch on the streets of Compton. An elderly grandmother had her necklace stolen by a young man who grabbed it and ran off.  There had been a rash of these types of thefts in Compton so we were asked to look at it and see what we could see.  We quickly found the incident within our data and saw the person approach the elderly lady and then quickly run off.  He ran about a block and got into a car.  We followed the car past a number of ground cameras and were able to get a description.  When we followed the thief backward we saw that he had gotten out of the same car which then circled the block to meet him on the other end.  We followed the car further backward in time to a gas station where he had the car and the person paying the clerk $5 for gas on video 20 minutes prior to the robbery.  All of this information was provided to the LA County Sheriff detectives we were working with.


Can you see car crashes and figure out who is at fault?

We typically see over 50-60 car crashes in a day.  We often will look at the hit and runs as they are more serious.  We can see many auto crashes and can often determine who is at fault.  We often see people involved in car accidents running multiple stop lights and stop signs prior to the actual accident.  One accident we saw the other day had a car run a red light and crashing into another car, the driver got out and carjacked another car at gunpoint and drove off.  We followed the car and shortly we had his image in another ground based camera about a mile away.  


When is the imagery available and how long does it take?

CSP imagery is captured and processed in about 5 seconds. The imagery is available in near real time to CSP Analysts at the Police Department Watch Center and the CSP analysis center.  We have at times caught up to criminals as the fled a crime scene and had them pulled over in real time.


That sounds like a cool job are you hiring?

Yes.  We are always looking for analysts and interns.


Where do you find your Analysts?

We hire locally through hiring agencies and directly.  We have internships and other programs available.  Please see our website for more information.


Do I need a Security Clearance to be an Analyst?

You do not need a security clearance to be an analyst but you must be a person in good standing.  We will conduct a background check and a drug screening.  Many of our analysts will go on to obtain their DoD security clearances and take jobs in the Defense industry after getting the required experience to qualify for a position with us.