Spillover and peer effect models network depiction. Click on nodes indicates students own GPA, and their peers' GPA for students. Clicking on red dots (courses) reveals their name. The line indicates the grade each student obtained in the course and the campus where the course was taken.

Spillover and peer effect models network depiction. Click on nodes indicates students own GPA, and their peers' GPA for students. Clicking on red dots (courses) reveals their name. The line indicates the grade each student obtained in the course and the campus where the course was taken.

Talk prepared as service

PHUDCFILY

This analysis showcases a one-mode implementation of Geographical Network Visualizations as discussed in SSEM. Here connections are a function of sharing at least one 4-year neighbor. The units in gold are public 2-year colleges. An important contribution of this map is that distances are actually travel or commuting times, retrieved via a handcrafted function also described in SSEM.

This analysis showcases a one-mode implementation of Geographical Network Visualizations as discussed in SSEM. The units in gold are public 2-year colleges. An important contribution of this map is that distances are actually travel or commuting times, retrieved via a handcrafted function also described in SSEM.

This analysis showcases a two-mode implementation of Geographical Network Visualizations as discussed in SSEM. The units in gold are public 2-year colleges and the units in magenta are 4-year institutions. An important contribution of this map is that distances are actually travel or commuting times, retrieved via a handcrafted function also described in SSEM.

Humane Boarders Inc. (Fronteras Compasivas)} in collaboration with the Pima County Office of the Medical Examiner in Arizona continue to make geolocated information of migrant deaths publicly available. Since 2001 they have updated these counts carefully documenting locations and cause of death, whenever possible, of over 3,900 migrant who died or were killed while crossing the Sonoran desert. These analyses were prepared for SSEM.

Example of multiple plot mapping, interactive version

This is an example of the function poly_map(...) where all specified features are plotted and Hawaii and Alaska as shifted.

Visual prepared for SSEM. Replication code available at https://cutt.ly/ZZpzTej

This plot aims to address the question, what indicators may be included in the model that do not suffer from place-based multicollinearity. In other words, how can we rely on machine learning to detect indicators that may impact the performance of other indicators given their redundancy in explaining outcome variation?

Because 31 states (plus D.C.) rely on federal adjusted gross income information (Federation of Tax Administrators, 2022) to calculate their own income taxes, such states have also foregone significant tax revenue due to these SLID costs. Given SLID’s recentness and almost unstudied status, this remarkable state’s cost has remained under- or not discussed. The purpose of this blog is to provide population-level estimates of SLID costs by States’ foregone amounts and States' saved amounts given their lack of reliance on adjusted gross incomes in the calculation of their income taxes. These analyses are relevant because if States leaders continue to have no visibility on this issue, they will remain unable to leverage it strategically or adapt it to support college affordability perhaps more effectively.

Because 32 states rely on federal adjusted gross income information (Federation of Tax Administrators, 2022) to calculate their own income taxes, such states also forego significant tax revenue due to these SLID costs. A cost that has remained under- or not discussed. The purpose of this post is to provide population-level estimates of SLID costs by Fiscal year, separated by Federal and States’ foregone amounts. These analyses are relevant because if States leaders continue to have no visibility on this issue, they will remain unable to leverage it strategically or adapt it to support college affordability perhaps more effectively.

The tenets of geography of advantage/disadvantage suggest that the geographical indicators selected may be highly correlated, that is, zones with high crime are likely to have high poverty levels, for example. This correlation, which is typically observed in studies modeling environmental factors (Li, et al., 2016), may affect the observed variable importance of the predictors. Following Li, et al., (2016) before model estimation, variable inclusion criteria relied on a Feature selection algorithm (Kursa & Rudnicki, 2010) to detect all non-redundant variables to predict SLID variation via machine learning—this process effectively addresses multicollinearity issues by identifying and easing the exclusion of redundant features. This non-redundant feature selection was implemented using the Boruta function, a Random Forest regression procedure. Boruta is a wrapper algorithm that subsets features, the Xs and Zs depicted in equation (2) and train a model using them to try to capture all the relevant indicators with respect to an outcome variable. As depicted by Kursa and Rudnicki, relevance is identified when there is a subset of attributes in the dataset among which a given indicator is not redundant when predicting the outcome of interest. Procedurally, Boruta duplicates the dataset, and shuffles the values in each column referring to these shuffled indicators as shadow features. Then, a Random Forest algorithm is used to learn whether the actual feature performs better than its randomly generated shadow. The Boruta implementation relied on 1000 iterations; however, the optimal result was consistently found after 12 iterations indicating that each attribute had relevance levels higher than their shadow attributes. In conclusion, Figure 6 shows that all the features discussed in the data and methods section (see Table 2) were detected as non-redundant predictors of SLID tax expenditures.

PHUDCFILY

This visualization shows all community colleges with at least one four-year neighbor weighted by the number of neighbors. The color scheme at the institution level ranks institutions based on their predicted net prices after accounting for institutional, place-based indicators, and spatial dependence issues (obtained from model (1) in Table 3. These predictions are deemed to more realistically inform students' expected net prices of attendance given the particular configurations of their local neighboring structures, and place based indicators, such as cost of living. The states' color scheme are the result of averaging all these predicted net prices within each state; hence the resulting color gradient also ranks states based on the average expected affordability of their community colleges. This map then reflects concentration of these structures and affordability prospects.

The Heterogeneous Landscape of College Affordability in the Community College Setting. This map identifies four-year institutions located within feasible commuting distances from community colleges in the contiguous United States in rural zones. The net price predicted values correspond to model (1) in Table 3.

The Heterogeneous Landscape of College Affordability in the Community College Setting. This map identifies four-year institutions located within feasible commuting distances from community colleges in the contiguous United States in non-rural zones. The net price predicted values correspond to model (1) in Table 3.

In this lecture we cover spatial dependence tenets and offer code and data to replicate all the procedures user R.

A valid and new alternative to kernel density estimation (smoothing).

This visualization documents all available 12,226 within state articulation transfer agreements as presented by CollegeTransfer.Net (http://www.collegetransfer.net/Search/. This interactive version of state to state analysis includes number of agreements per state and the lines indicate the number of agreements within states. Institution data Source: IPEDS's NCES https://nces.ed.gov/ipeds/datacenter/

This visualization documents all available 1,696 out-of-state articulation transfer agreements as presented by CollegeTransfer.Net (http://www.collegetransfer.net/Search/. This interactive version of state to state analysis includes number of agreements per state and the lines indicate the number of agreements between states. Institution data Source: IPEDS's NCES https://nces.ed.gov/ipeds/datacenter/

This representation accounts for the prevalence of out-of-state articulation agreements analyzed at the program level as presented by CollegeTransfer.Net (http://www.collegetransfer.net/Search/. This interactive version includes number of agreements per program and the lines number of agreements between programs.

This representation accounts for the prevalence of within state articulation agreements analyzed at the program level as presented by CollegeTransfer.Net (http://www.collegetransfer.net/Search/. This interactive version includes number of agreements per program and the lines number of agreements between programs.

This visualization documents all available 12,226 within state transfer agreements as presented by CollegeTransfer.Net (http://www.collegetransfer.net/Search/. This interactive version includes number of agreements per institution and the lines indicate distances and number of agreements between dyads. Institution data Source: IPEDS's NCES https://nces.ed.gov/ipeds/datacenter/

There are 1,696 transfer agreements between CCs and public and private not-for-profit four-year institutions. What are the main drivers behind these connections? What are the most frequent programs? What can we learn from this potential mobility? Data sources: CollegeTransfer.Net (http://www.collegetransfer.net/Search/ IPEDS's NCES https://nces.ed.gov/ipeds/datacenter/"

This lecture covers mapping and visualization fundamentals using tigris and tmap, along with data obtained from the ACS, IPEDS, and IRS.

Steps required to create matrices

This visualization documents the agreements between community colleges and four-year institutions taking place in North Carolina. Data sources: CollegeTransfer.Net (http://www.collegetransfer.net/Search/ IPEDS's NCES https://nces.ed.gov/ipeds/datacenter/

Data formats and sources Spatial Socioeconometric Modeling (SSEM)

Percent of Balance 90+ Days Delinquent by Loan Type Source: New York Fed Consumer Credit Panel/Equifax https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/xls/HHD_C_Report_2019Q4.xlsx Dynarski (2016) showed that only about 18 percent of students who owe more than $100,000 default while 34 percent of those owing less than $5,000 do so. After trying to identify these students in National Postsecondary Student Aid Study (NPSAS), I realized we do not have data that allow us to identify them. Accordingly, to search for strategies that address questions like "how to protect student loan borrowers who accrued lower amounts from defaulting on their loans?" we first need to learn more about them. Dynarski, S. (2016). The trouble with student loans? low earnings, not high debt. Brookings Note. Available https://www.brookings.edu/research/the-trouble-with-student-loans-low-earnings-not-high-debt/

This network represents 116 participants coming from 75 institutions across the United States and abroad.

Even if this is just descriptive, if you live in a county where the vast majority of confirmed cases has died, I would recommend seeking treatment elsewhere.

Descriptive analyses of COVID cases and deaths per county based on New York Time data (https://github.com/nytimes/covid-19-data). 1st layer: No. of cases per 1000 county inhabitants 2nd layer: No. of deaths per 1000 county inhabitants 3rd layer: Percentage of deaths per cases within county Code available here: https://drive.google.com/open?id=1x_BpCZ4ibiqr6tOiq2PRGsHHQYtRvpWC&authuser=msgc@email.arizona.edu&usp=drive_fs

Measuring Tuition and Fee Affordability Given Local Configuration of College Options: Community College Students’ Real Cost Based on Four-Year Nearby Options A neighbor is located withing 40 miles

A neighbor or link is identified if a community college and a four-year institution are located within 20 miles in non-rural areas.

Why if we know that the for-profit sector is the most expensive, loan wise, we keep having it?

Senators Elizabeth Warren and Bernie Sanders continue to push their student loan debt-free agendas as part of their political campaigns. Sanders's plan promises to forget all debt, while Warren's plan is a tiny bit more conservative by planning to forget about 95% of all outstanding debt. How realistic are these promises when contextualized against other forms of debt growth over time? In January of 2003, forgiving student loans would have equated to forgiving all House Equity amounts (about $.24 trillion). In October of 2009, it would have equated to forgive all outstanding auto OR home equity loans. Three months after, it would have equated to forgive all combined outstanding amounts of credit card loans. Finally, the most recent data indicates that forgiving student loans would equate to forgiving the COMBINED home equity and credit card amounts.

Placebo test using a false "crisis" start date to test whether crisis is actually driving debt accrual in the United States

Are economic crises driving student loan debt accumulation?

During this same period the for-profit sector also accounted for 60 percent of the net total closure, with 1,635 closings. The public and private not-for-profit sectors closed 257 and 818 campuses, respectively.

The for-profit opening rate is also the highest across all sectors. A longitudinal analysis of IPEDS data, shows that between 2000 and 2018 a total of 1,984 for-profit institutions, including campuses, started operation. This number represents 73 percent of all openings, with 217 and 318 public and private not-for-profit campuses, respectively, opening over the same period.

The for-profit closure rate is the highest across all sectors. A longitudinal analysis of IPEDS data, shows that between 2000 and 2018 a total of 3,084 for-profit institutions, including campuses, have stopped operation. This number represents 65.7 percent of all closures, with 474 and 1,136 public and private not-for-profit campuses, respectively, closing over the same period.

Student loans are the only form of debt that (a) did not slow down during the most recent economic crisis and (b) have grown 6.25 times since 2003, reaching $1.6 trillion as of December 2019.