A2-KenghaoChang
From CS294-10 Visualization Fa07
Contents |
[edit] Introduction
I take advantage of this assignment to analyze the interview results collected from a non-profit community center. My advisor and I owed them the analytic results. The center's mission is to support and strengthen the community by serving youth and families through counseling, housing, education, and advocacy. The questions in the interview basically are to understand the mental status (e.g. anger management) and their home background (e.g. financial problems) of the runaway and homeless youths.
[edit] Dataset 1: interview results
For each youth individual, interviewers asked nine questions and each question is to be answered on the scale 1-4.
| Question index | scale: 4 | scale: 3 | scale: 2 | scale: 1 |
| 1 | Expresses anger appropriately. | Sometimes loses control when angry. | Frequent loss of control when angry; expression of anger may include destruction of property. | Cannot control angry behavior; expression of anger is violent toward others. |
| 2 | Copes well with problems that arise on an everyday basis. | Struggles with problems that arise, but seeks assistance from others.Only occasional crisis stage. | Problems often reach crisis stage but are dealt with eventually; not be able to express feelings. | Always in a state of crisis. Has no coping mechanisms. |
| 3 | Fair and consistent with rewards and punishments. Coping well with parental responsibilities | Mostly consistent in rewards and punishments. Coping somewhat well with responsibilities. | Having difficulty with parental responsibility; discipline is somewhat inconsistent. | Discipline is poor; inconsistent; lack of limit setting; not coping with parental responsibilities. |
| 4 | Has support system he/she can count on for emotional support and concrete help during stressful times. | |||
| 5 | Has some supports ut no one to turn to when they are unavailable. May express desire to have more friends or other support. | Has some supports ut no one to turn to when they are unavailable. May express desire to have more friends or other support. | Has support network in other cities or counties, but only acquaintances here. Generally feels isolated. | Family/Couple never does anything together.Has no contact other than extremely casual. Doesn't know how to make friends or access community resources. Feels very isolated. |
| 6 | Abstinence from illegal drugs and only occasional use of alcohol. | Drinks or use drugs occasionally. Substance use does not affect everyday activities. | Uses variety of drugs and/or alcohol heavily. Use has affected job/school and/or family relationships. | Dependent on alcohol or illegal drugs. Primary focus is on getting and using drugs. |
| 7 | No gang activity. | Peripheral gang activity; is a "wannabe"; not an official gang member. | Frequent involvement in gang activities. | Gang is primary affiliation or identity. |
| 8 | Individual/family has full-time employment and/or receives adequate child support. | Has part-time employment and/or receives inadequate child support payments. | Relies consistently on public assistance and/or is infrequently employed. | No consistent, legal sources of income. |
| 9 | Can live independently within means (AFDC, employment, SSI); able to handle emergencies. | Living somewhat "on the edge"; crisis can be imminent. | Financial crisis occurs monthly. | Consistently living beyond budget; in constant state of financial crisis. |
There are around 650 youths answering the nine questions.
[edit] Dataset 2: number of bednights
In general, the runnaway youths were sent to the center by social workers. The youths were asked to participate the conselling program in the center and workers in the center would ask them the nine questions. During the course of the conselling program, the youths would be housed for several bednights and the number of housed bednights would be recorded. There is a single record per bednight for each youth. Each record is only dated with the year of the bednight.
Since there is some privacy concern of the data, I'm not revealing the name of this center here and i'm not posting the data publicly here.
[edit] Questions being answered
The basic questions I am trying to answer from the data set is: how's the average reponse by the youths to the nine questions? is there any correlation between the mental status, financial issues, drug use, and the family background of runaway youths? In addition, how's the load to house youths in the center?
[edit] Analysis
[edit] Q1 - What's the general response to each question?
First of all, i would like to know how the runaway youths responded to the nine questions in general . The following charts represent the distribution how the youths responded to each of the nine questions in the scale 1 to 4 (dk: don't know; na: not applicable). We plotted the charts only from baseline results.
Based on the bar charts, most of the youths frequntly loss control when angry (Q1), sometimes struggle with problems (Q2), don't have support system they can count on for emotional support (Q4), generally feel isolated because of lacking support networks(Q5), heavily use drugs and achohol (Q6), and frequently involve in gang activities (Q7). Nonetheless, Q3, Q8, and Q9 cannot be analyzed because the majority of the answers fall into dk (don't know) and na (not applicable).
[edit] Q2 - Is there any correlation between social support and aquired problems?
The second question I'm interested is, can we find correlation from the interview results? If there is little emotional support a youth can count on, would that cause he/she manage anger badly? The same, if there is little emotional support, would the youth be addicted to alchohol and drug?
To visualize whether there is correlation between any two questions, I chose "Heat Map", which is provided by Tableau. In our case, we want to visualize 3-dimentional data, and the heat map is capable to visualize it to some extent. We want to visualize whether there are more youths responding the given two questions in high scale value at the same time. If the answer is yes, there should be certain level of correlation.
Given two questions Qx and Qy, the heat map is a 4x4 matrix. Each row i represents a scale value for question Qx and each column j represents a scale value for question Qy. The heat map uses the size of a rectangle to represent how many youths answering Qx in scale i and Qy in scale j. The bigger the rectangle, more youths answering the two questions with scales falling in that cell.
Here, I picked three combinations Q1(anger managment capability)-Q4(emotional support level), Q4(emotional support level)-Q6(achohol and drug usage level) and Q5(social support level)-Q7(gang activities evolvement level). The visualization shows that there are more youths answering the given two questions in high scale value.
[edit] Q3- How is the housing load of the center?
I'm wondering how much load the center had, in order to accommodate the runnaway youths? Because of the nature of the dataset that there is a single record per bednight for each youth and each record is only dated with the year of the bednight, i sum up the total number of bednights for each year.
From the bar chart, in year 2001-2004, the load is similar, but the load is much higher than the load in year 2000 and 2005. Either we can interpret that in years 2001-2004, there are more runnaway youths that year 2000 and 2005, or we can suspect that the data was not collected completely in the full year of 2000 and 2005; it might be collected only in last half part of 2000 and only in the first half part of 2005.
